The British Library Board 1995
The opinions expressed in this report are those of the authors and not necessarily those of the British Library.
RDD/G/260
Significant further development and investigation is nevertheless required. A proposed demonstrator project - based around Scottish University Libraries and the BUBL Subject Tree initiative, but sufficiently 'open' to encompass other sites and approaches - is both feasible and essential, and would provide a focus for Z39.50 developments in the UK.
Z39.50 clients and associated Z39.50 OPACs describing resources could become preferred network navigation tools with other specific NIDR client types (WWW, gopher, WAIS, others) loaded as required. Library involvement is essential to sustainable Internet cataloguing initiatives.
1. Aims of the Study and the
Problem in Outline
2. The Proposed Solution: Assumptions
Underlying the CATRIONA Model
3. The Proposed Solution: Outline
Description of the CATRIONA Model
4. Work Done, Problems Encountered,
Limitations of the Project
5. Feasibility of the Model and of a
Follow-up Demonstrator Project
6. Summary of Development
Requirements
7. CATRIONA Demonstrator: General
Points
8. CATRIONA Demonstrator: Specific
Phase II Proposals
9. CATRIONA Phase II Requirements: Skills, Personnel,
Organisation, Costs
10. CATRIONA and the BUBL Subject
Tree Project: Further Information
A Z39.50 GUI OPAC client available from one library system vendor (GeoPac Release 1.23 from Geac) has been observed to be capable of
Since the client in question can be set up so as to add additional Z39.50 servers to its search list, it would clearly be possible to create a distributed catalogue of Internet resources, to access that catalogue on a server by server basis, and use it to find and display networked electronic resources. Since it is also possible to use the client to search standard Z39.50 OPACs, the process is integrated, as envisaged within the CATRIONA model, with that of retrieval of information on traditional hard copy resources, and with that of finding and accessing locally-held or subscribed-to electronic information resources.
At least one other library system vendor appears to have a Z39.50 client with similar capabilities. The OPAC client in the Dynix Horizon system (Version 3.2) has been shown to have facilities which should enable it to load Netscape in response to a USMARC record containing a URL in the 958 field and of retrieving and displaying an electronic information object (EIO) located at a remote site, although this has not actually been observed working within the CATRIONA project. Dynix do not yet utilise the 856$u subfield, but are expected to make changes in the near future. Since the client is capable of conducting a search of remote Z39.50 OPACs and retrieving records with URLs in 856$u, it should have similar capabilities to the Geac client once the field utilised for the URL has been changed. A useful feature of the Dynix client not at present found in GeoPac is its use of a .ini file to specify the EIO viewers. This should allow a wide range of different 'viewers' to be specified for a range of different EIOs and is therefore more flexible and potentially more powerful than the present GeoPac approach.
It seems possible that a client of this kind - a Z39.50 client able to load a range of different NIDR (Network Information Discovery and Retrieval) clients in response to specific types of URL, and backed up by a range of Z39.50 catalogues describing electronic resources - would be a serious candidate to become a key network navigation tool, perhaps in preference to other specific NIDR client types (WWW or WAIS or Gopher clients) although such clients might well be further developed to encompass Z39.50 compliance.
Although it would be entirely possible to implement a basic demonstrator project without it, further software development is required to create a more sophisticated user-friendly environment for searching a distributed catalogue of the Internet and retrieving/viewing electronic resources envisaged in the full CATRIONA model. An initial list of required features and facilities has been compiled, although it is probably the case that it will only be possible to clearly identify and describe some of the details relating to the refinement of the model itself, and to associated software development requirements, within the context of an actual demonstrator project. Another aim of the demonstrator project would be to test whether the creation of a libraries-based world-wide distributed catalogue of Internet resources constructed from regional building blocks, 'Union' OPACs linked to cooperative cataloguing projects, subject OPACs, and other non-library elements is a practical proposition. Such a project would also help provide a focus for Z39.50 development, implementation and experimentation in the UK.
There is good reason to suppose that a demonstrator based on the full model is feasible. A range of Z39.50 clients have been examined within the project, and whilst none of these has all of the facilities required for the full model, many of these facilities have been seen to work in one client or another. Other requirements for implementing a demonstrator based on the full model are also in place, or will be in future. For example, on the co-operative cataloguing/'Union' catalogue front, it is now clear that both OCLC and RLG will be creating USMARC records describing EIOs (Electronic Information Objects) , and it seems probable that co-operative cataloguing of Internet resources will follow these initiatives; indeed the OCLC project entails a co-operative element as do the RLG plans (OCLC, 1994), (Washburn, 1995) .
The requirements for a Phase II demonstrator project have been sketched out and costed, although it is true to say that a range of Phase II options exist and that the project could, within certain limits, be tailored to fit a desired level of funding. The position of the commercial companies involved is subject to negotiation regarding the nature of the follow up project. The estimated Phase II cost of 470,000 pounds over 3 years is provided for guidance only, subject to further discussion of the exact nature, shape and timescale of the demonstrator project.
A Phase II project would aim to develop a Scottish distributed network resources OPAC, integrated with existing traditional OPACs (i.e. records in the same catalogue) and accessible through a series of enhanced traditional OPAC clients produced by a number of different suppliers. The end result will be an embryonic but stable operational (non-experimental) system based on the CATRIONA Phase I model - i.e. a system and associated infrastructure capable of supporting a sustainable, distributed and scalable approach to network resource discovery, description, retrieval, and presentation/utilisation, based upon existing library procedures, practices and standards (e.g. Z39.50, MARC). The system would also incorporate library system based access control mechanisms. The system would be 'open' in that other libraries, organisations and suppliers could join at any time, simply by adopting the same, or a similar, model, and it is by this means that more comprehensive library-related catalogues of networked resources might eventually emerge. The possibility that some of the proposed FIGIT subject services should also be based on MARC and Z39.50 should be considered.
The current version of the CATRIONA model (Version 4) is described in Section 3 of the report, with further illustrative detail in Appendix B.
Further information on the CATRIONA feasibility study, and on other related information and projects, can be found on the BUBL Information Service at URLs:
CATRIONA Web Site
CATRIONA Gopher Site
___________________________________________________________________________
The CATRIONA feasibility study proposal arose out of a desire to tackle, from a library service perspective, four related elements of what is essentially the same problem:
1.1. Providing library users with a reliable and user-friendly means of finding, retrieving or otherwise utilising the growing range of 'Electronic Information Objects' (EIOs) - electronic texts and journals, images, sounds, interactive services, multimedia documents, programs, video, human expertise, and so on - available over the Internet. One that utilises library- standard finding tools and library-standard descriptions and which preferably also encompasses access to information on traditional (i.e. non-electronic) resources.
1.2. The need identified by systems and reference librarians to provide library and campus network users with seamless access, integrated within a single workstation (ideally, of various types: PC, MAC, X-Windows), and mediated through the library OPAC, to a range of electronic information resources held, or subscribed-to, locally, including:
This ideally to extend beyond these elements to encompass both Internet access as described at 1.1 above and, potentially, access to local campus-based electronic resources and to the wider 'virtual library' as outlined at 1.3 and 1.4 below.
1.3. The question of how best to ensure that the growing range of e-resources available at an institution is exploited to its fullest both across campus at the institution itself and, where so desired for income-related or promotional reasons, beyond the local institution via network access - the suggestion being that the best way of doing this may be to catalogue these resources in the network-accessible library OPAC.
1.4. The need to find a practical mechanism for finding, retrieving and 'viewing' or utilising EIOs in the much-heralded distributed 'virtual library', which will simply extend and encompass what is already available over the Internet and which, like the Internet itself, can ultimately only function if it has at its heart a system and associated infrastructure capable of supporting a sustainable, distributed and scalable approach to networked resource discovery, description, retrieval, and presentation.
__________________________________________________________________________
Chapter 2. The Proposed Solution:
In proposing the CATRIONA model (described in Section 3 and
Appendix B of this report) as a means of dealing with this
problem, a number of assumptions have been made about the
kind of solution required. These are presented here for three
reasons:
Assumptions
Underlying the CATRIONA Model
Many of the assumptions were identified or brought into clearer focus during the project itself.
Assumptions:
2.1. Any solution proposed must be practical, sustainable, scalable and appropriate to the problems faced by library and campus-wide users in any given local context. That is, it must:
2.2. A searchable catalogue (or integrated group of catalogues) of Internet resources containing library-standard resource descriptions is what is required. Finding tools which rely on general subject headings, hierarchical menus, and webs of hypertext links of world-wide scope, and indexes based on menu-entry descriptions are no substitute for full descriptive cataloguing and library-OPAC level search engines. They would not be considered adequate in a single library and certainly cannot be regarded as an adequate answer to dealing with the world-wide resources of the Internet. This is not to suggest that non-library descriptive cataloguing projects based on e.g. the Text Encoding Initiative (TEI) Project (Sperberg- McQueen and Burnard, 1994) do not also have a significant role to play.
2.3. It cannot and will not be done on the scale, or to the level of reliability, descriptive detail, and sustainability required, by volunteers (e.g. BUBL and other subject trees), resource creators, individual services, or publishing companies, although such efforts will continue to play a role in the overall process. This is true in the world of printed resources and there is no evidence to suggest that the world of electronic resources will be significantly different in this respect. Even TEI headers, descriptive metadata for SGML (Standard Generalised Markup Language) documents which, it may be assumed, will be created by authors as an integral part of the electronic document itself, (Giordano, 1994a) are recognised as being limited and likely to require expansion and extension to make them suitable to (local) library purposes. Description of electronic resources held, created, or utilised locally must become a recognised part of the cataloguer's official function and be carried out on a world-wide scale. None of the other methods mentioned above will be sufficient to solve the problem, nor does there seem to be any evidence at present to suggest that automated methods based on resource content can replace the need for descriptive cataloguing.
2.4. The library community has the skills, experience, traditions and organisation required to enable them to solve the various aspects of this problem. They have already begun to respond to the need to catalogue local electronic resources such as CD- ROMs and it is possible that, for purely local reasons, most Internet EIOs important to library users will be catalogued locally. At a basic level, this is all that is required to ensure that the whole Internet is catalogued to a level sufficient for library purposes, and the other types of effort mentioned above should cover most of what is not covered by this means.
2.5. The established practice in the library community of using and contributing to co-operative cataloguing ventures can play a significant role in reducing the cost and burden of cross- Internet cataloguing.
2.6. One big OPAC for the Internet is not a practical proposition for a whole range of organisational, financial, cross-national, and language-related reasons. Apart from problems of scale and of keeping up with demand and publishing effort and of deciding what should be catalogued and in what order of priority, there is also the question of who would fund it and run it. Would RLG (Research Libraries Group) let OCLC (Online Computer Library Center) go it alone for example, or would they set up in competition? Would one country allow another to be the only host? A joint approach by all organisations and countries might be a possibility, but how many years would it take them to reach an agreement? And what about the problems caused by different languages? Given these kinds of problems, would a centrally funded joint project ever become a reality?
The idea of a single central OPAC (or even several) also has other drawbacks. It could not, by its very nature, be a catalogue of local electronic resources, which means that it would not solve the aspects of the total problem identified at 1.2, 1.3 and 1.4 above, nor would it cover Internet resources authored locally and not more widely publicised. A single central OPAC must also solve the problem of the rapid identification and updating of changed URLs based at remote sites. Although the Internet community is presently investigating ways of dealing with this problem, it is arguably much easier to solve at present if both the catalogue record maintenance and the URL (Uniform Resource Locator) of the resource itself are under local control.
For all of these reasons, it is assumed that the best solution to the problem is not one or several large central OPACs, but a distributed catalogue of Internet resources based on a range of regional or other building blocks, and that the way forward is to design a mechanism and set of protocols and procedures which will allow any library (or other group) to participate and which will allow the whole catalogue to be built up gradually and ultimately be as comprehensive as possible.
2.7. A distributed search of every library and all other OPACs on the Internet is not efficient either in terms of network use or in terms of user (and library workstation) time, so 'Union' catalogues created through co-operative cataloguing ventures should be part of any efficient system - especially since, if these catalogues exist anyway, users will use them if they provide faster resource discovery.
2.8. A distributed catalogue able to be built up gradually as new catalogues 'join' the system MUST be based on standards, especially Z39.50 but also accepted record formats such as MARC (Machine Readable Cataloguing) . It must also, of necessity, be an 'open' system, which is to say hospitable to the products of all systems suppliers and, hence, potentially, to all possible library and other participants. (See 3.1.1 for reasons for choosing MARC).
2.9. Even after the proposed distributed catalogue is well- developed and established, other types of Internet metadata - IAFA (Internet Anonymous FTP Archives) templates (Deutsch et al, 1994), TEI headers (Sperberg-McQueen and Burnard, 1994) and searching tools - e.g. Veronica, ICE - will nevertheless continue to be important and necessary for some considerable time, perhaps indefinitely. Any system must therefore encompass the use of such tools. There will be an associated need for reliable regional or national indexing services - Veronica, Archie etc. - presumably funded in the UK by the Joint Information Systems Committee of the Higher Education Funding Councils for England, Scotland and Wales (JISC).
2.10. Since paper-based resources will continue to be important for some time to come, there is a need to integrate searching for electronic resources with hard-copy searches and related paper-based ILL and document delivery services, hence the need to integrate with local OPACs and involve library system suppliers.
2.11. Libraries will wish to offer users access to most Internet and local network resources but they will wish to manage access for a number of reasons (See 3.4 below). Any system must therefore encompass access control mechanisms.
2.12. Access control can be achieved via library system user classification and verification plus control over which remote catalogues, other sources of metadata and so on are catalogued locally
2.13. A client-server model is essential if resources such as images, sounds, video etc. are to be retrieved and displayed or 'played', but VTxxx access is also necessary despite limitations (i.e. text-only) because of limitations on technical capabilities at individual sites and on individual desks.
2.14. The need to cope with resources only accessible through different NIDR clients, such as WWW (World Wide Web) or WAIS (Wide Area Information Servers) clients and with other as yet undeveloped methods of dealing with Internet resources requires that the CATRIONA client must be able to load a range of these and to distinguish between them as appropriate.
2.15. In the longer term, the library community world-wide will build up an information and communication network which will enable them to keep in touch with new publishing developments on the Internet in rather the same fashion as they now do with hardcopy resources. Until this network is established, however, projects like the BUBL Subject Tree Project, with its associated new resource finding and monitoring efforts, will continue to be important if previously unknown EIOs and new editions of remotely produced resources held and catalogued locally are to be identified and catalogued (See Appendix I).
___________________________________________________________________________
Chapter 3. The Proposed Solution:
The original CATRIONA model as specified in the funding
application documentation (available at URL:
http://www.bubl.bath.ac.uk/BUBL/catriona.html) was developed
and clarified as the study progressed. What is described here
encompasses these changes. Some of the details were added
after the results of the tests of the feasibility of the model
described in Section 5 of the report were known. It is expected
that further development and clarification will take place,
particularly in the context of the proposed demonstrator project,
which is seen as an essential pre-requisite of refining the model
in a practical context, but also as a result of continuing
discussions with various interested parties and of continuing
investigations of related projects.
Outline Description of the CATRIONA Model
The following is a list of the main features of the present version of the CATRIONA model:
In both cases, items catalogued will include a wide range of resource types - electronic text, multi-media documents, images, sounds, programs, videos, interactive services, other OPACs etc. In both cases, the records describing the resources utilise the MARC standard enhanced as recommended by MARBI (MARBI, 1993a, 1993d, 1994a). (Note: These enhancements were developed initially for USMARC and implementation in other MARC formats is limited as yet. See Appendix E - Dutch InfoServices Project). This means, amongst other things, that they hold in 856$u a URL or URLs for each electronic resource catalogued, information that can be utilised by clients such as Mosaic or Netscape to locate, retrieve and display or 'play' the resource on a local workstation.
The use of MARC is proposed for a number of reasons:
Since the 'playing' of such things as images, sounds, video and programs requires facilities only available on a microcomputer or workstation as opposed to a terminal, a client-server model is assumed, with at least the presentation or display level under the full control of a GUI (Graphical User Interface) client on a microcomputer or workstation. However, it is recognised that sites and individuals with more limited technical facilities must also be catered for and that a VTxxx client capability with more limited facilities for retrieving and displaying electronic resources (i.e. text-only) is also essential.
The model also recognises the need for a mechanism to enable URLs to be updated in 'Union' catalogues - OCLC, CURL (Consortium of University Research Libraries) etc.
One advisor to the project has suggested that this approach does not deal adequately with the situation where the original of the resource held locally is changed at the remote site and that the URN (Uniform Resource Name) to URL resolution service proposed by the Internet Engineering Task Force (IETF) is needed to resolve this problem (Daniel, 1994b), (Mitra et al, 1994). Most experts in this area also appear to take the view that the URN to URL resolution service will offer the basis of a better solution to the problem of broken URLs than the one proposed at present in the CATRIONA Project. For these reasons, it is felt to be essential that the demonstrator project be so designed as to be adaptable to developments in this area. Nevertheless, the following points are felt to be worth noting:
It is assumed that reliable and important services of this kind will be catalogued in the local and other OPACs on the Internet in the same way as described at 3.3 above and that information on them and how to access them will be 'served up' to the user whenever the circumstances of the search make them appropriate. It is further assumed there will be a role for JISC- funded services like BUBL, NISS (National Information on Services and Systems) and others in the provision of stable and reliable access to services like Veronica, Archie and WWWW.
It is assumed that access control can be managed via a combination of user classifications on the local library housekeeping system and selective cataloguing and judicious classification/subject indexing of other Internet catalogues and other Internet indexing services. More work is required in the practical context of a demonstrator project to clarify exact requirements and identify feasible solutions. It is possible that there will be a requirement to develop functionality based on some aspects of Z39.50 Version 2 not yet addressed by library systems suppliers (Access Control and Resource Control).
The model has changed and developed throughout the project. A description of Version 4 of the model illustrating how it would interact with the user in various situations is presented in Appendix B, together with the development requirements appropriate to each stage.
___________________________________________________________________________
These limitations on the time available for the project have meant that, whilst all or most of the various topics germane to the enquiry have been addressed to some extent, only core concerns relating to the feasibility of the model have been investigated in any detail. Another consequence is that the report is necessarily a snapshot of progress at a particular point in time. More work needs to be done on the wider field of topics and projects and this is one of several reasons to consider a follow-up demonstrator project.
4.1. Obtaining feedback from the library community, the wider Internet community, and, to a lesser extent, from system suppliers and similar organisations, proved more difficult than anticipated. Some of this was identified as arising from difficulty in understanding the model and a great deal of unplanned work was done to clarify the written description both through re-writing it and through addressing the question at various meetings.
4.2. When the project became public knowledge a large number of library system suppliers not originally involved in the project indicated that the felt they should have been involved. For a number of reasons, it seemed wise to involve them, and doing so proved useful on a number of counts, but it also inevitably increased the work that had to be carried out within the project.
4.3. The need to install and look at an associated number of Z39.50 clients, and the inevitable extension of technical difficulties that this caused, also increased the work that had to be carried out, but also helped identify and assess the feasibility of CATRIONA Phase II development requirements.
4.4. A key assumption - that the client required for CATRIONA would be created by enhancing the internal functionality of existing OPAC clients proved to be misconceived. It became clear that the best approach to this was for such clients to be able to call up an appropriate NIDR client (e.g. Mosaic, or a WAIS client), that some suppliers were already moving in that direction and others planned to do so. This meant that it became sensible to concentrate on the features of existing Z39.50 clients rather than on those in NIDR clients as originally planned. It also meant that suppliers felt no need to discuss the subject of NIDR client capabilities with NIDR client developers as originally planned.
4.5. The discovery of a Z39.50 client able to act on a URL in a USMARC record meant it was sensible to attempt to find a Z39.50 site with records of this kind in their database and to set up the client so that it could inter-operate, again a time- consuming process. A message requesting information about Z39.50 servers with URLs in 856$u was sent to the USMARC, Z3950IW and GILS lists (See References). Only one of the responses contained information which was sufficient to make a Z39.50 connection to the server and access the records. Each of the other responses required further follow up for further details of site address, search type required and encoding of the 856$u subfield.
Creation of a wide range of hypertext references on the BUBL
WWW and Gopher servers:
Available at the following URLs:
http://www.bubl.bath.ac.uk/BUBL/maincatriona.html
gopher://bubl.bath.ac.uk:7070/11/Link/Catriona
Development of CATRIONA model Versions 2, 3 and 4
(Version 1 in original bid. Available at
http://www.bubl.bath.ac.uk/BUBL/catriona.html).
Discussion of model over e-mail and at face-to-face meetings with systems suppliers: Geac, Dynix, FDG, SLS, SCG, MDIS. And with cataloguing services: CURL, OCLC, RLG.
Dissemination of information relating to the project through
BUBL and e-mail discussion lists.
(LIS-LINK, LIS-BAILER, Z3950PIG etc.).
Presentations of the model and cataloguing issues as well as panel discussion at the Cataloguing and Indexing Group Scotland meeting (See Appendix K).
Discussion of Z39.50 issues and attendance at a number of related meetings (See Appendix H).
Survey, installation and examination of a number of existing Z39.50 OPAC clients: GeoPac Release 1.23, WinPAC Beta 1.0, Horizon 3.2, VIZION, Telnet access to Libertas 6.3 and Innopac clients. (Telnet provides a connection to a remote computer over the Internet or a remote login). Observation of Oracle Libraries client and MDIS Lion prototype client during visits.
Survey, installation and examination of a number of NIDR clients: Cello, MS windows Mosaic, Netscape, Hgopher, WinWais and others.
Testing the feasibility of the CATRIONA model at a basic level.
Design of proposed demonstrator project.
Formulation of development requirements for proposed demonstrator project.
Discussion of likely costs of CATRIONA Phase II with selected suppliers.
Preparation of report
Further information is provided in Appendix C under the above headings.
___________________________________________________________________________
Chapter 5. Feasibility of the Model
The stated aim of the CATRIONA feasibility study was to
investigate the technical, organisational and financial
requirements for the development of applications programs and
procedures to enable the cataloguing, classification, and
retrieval of documents and other resources over networks, and
to explore the feasibility of a library system supplier led
collaborative project to develop such applications and
procedures and integrate them with one or more existing library
housekeeping systems and associated OPAC interfaces. A
particular model for cataloguing network resources over the
whole Internet was proposed (For Version 1 of this model see
URL: http://www.bubl.bath.ac.uk/BUBL/catriona.html). This
assumed a distributed catalogue of Internet resources with
distributed cataloguing, local control of record updates and EIO
location, and a key role for co-operative cataloguing and 'Union'
catalogues through organisations such as OCLC, CURL and
BLCMP. Initially, it had been assumed that the feasibility of the
model and the proposed demonstrator project would be
examined through critical analysis and discussion of the model
and of technical, cataloguing, and software development
requirements with library system suppliers, NIDR client
developers, systems and reference librarians and other groups
with interests in this area. However, as the study progressed it
became evident:
and of a
Follow-up Demonstrator Project
This raised the possibility that a more practical test of the feasibility of a demonstrator project based on the model could be conducted. GeoPac was installed on an IBM compatible PC running a Winsock compliant version of PCNFS 5 and shown initially to be capable of retrieving a catalogue record on a database on the same machine, responding to a local URL by offering the option to link to the image described in the record, then loading an image viewer to display the locally held image. The next step was to install Mosaic for MS Windows on the machine and set up GeoPac so that it could locate and load Mosaic as required. This done, it was necessary to identify one or more remote Z39.50 OPACs describing electronic resources and containing URLs. Messages were sent to a number of electronic mail discussion lists (USMARC, Z3950IW and GILS) and three OPACs - at Brigham Young University, Butler University and North Carolina State University - containing URLs were identified. Z39.50 connection details were obtained and GeoPac was set up so as to be able to access the OPACs. Details of the records themselves and how to search for them were also obtained and an attempt was made to connect to each of the databases in turn to retrieve the enhanced USMARC records and get GeoPac to load Mosaic and retrieve and display the described EIOs.
The first database searched was the BYU server at Brigham Young University, in which there were two records that have been enhanced to include URLs in 856$u. On retrieval of the records, the client's 'link' button was highlighted and clicking on this had the effect of making the OPAC client load a suitable 'viewer', and pass it the URL so that it could locate, retrieve and display the EIO described by the MARC record. The two records from the BYU server were retrieved in turn and the 'link' button 'clicked on' as described above. GeoPac duly retrieved the EIO described, loading an appropriate viewer to display it. In the first case, the viewer loaded was the WWW client Mosaic and the EIO was a full text article on a gopher server in Minneapolis. In the second case the viewer loaded was a .gif image viewer and the EIO was a .gif file located at Brigham Young University. The second database searched was at Butler University Library which has at least 3 records with URLs in the 856$u subfield. The viewer loaded was Mosaic and the EIO was a full-text HTML format file located at Indianapolis. The third database searched was at North Carolina State University Library, where there are records with URLs in the 856 field, but not in the $u subfield. In this case the EIO could not be retrieved.
Subsequent tests showed that GeoPac could also load Netscape in preference to Mosaic, although in its present incarnation it can be configured only to run one NIDR client at a time. (Note: it can also open a Telnet session in addition to this and utilise an image viewer if required). The client was subsequently installed on another PC which utilised Beame and Whiteside's Winsock compliant TCP/IP in preference to PCNFS V.5. The tests were repeated with similar results.
These tests and subsequent investigations show that the idea of a distributed catalogue of Internet resources integrated with standard Z39.50 library system OPAC interfaces is already a practical proposition at its most basic level. Since GeoPac can be set up so as to add additional Z39.50 servers to its search list, it would clearly be possible to create a distributed catalogue of Internet resources, to access that catalogue on a server by server basis, and use it to find and display networked electronic resources. Since it is also possible to use the client to search standard Z39.50 OPACs, the process is integrated, as envisaged within the CATRIONA model, with that of retrieval of information on traditional hard copy resources, and with that of finding and accessing locally-held or subscribed-to electronic information resources.
At least one other library system vendor appears to have a Z39.50 client with similar capabilities. The OPAC client in the Dynix Horizon system has been shown to have facilities which should enable it to load Netscape in response to a USMARC record containing a URL in the 958 field and to retrieve and display an EIO located at a remote site. Dynix do not yet utilise the 856$u subfield, but plan to make changes in the near future. Since the client is capable of conducting a search of remote Z39.50 OPACs and retrieving records with URLs in 856$u, it should have similar capabilities to the Geac client once the field utilised for the URL has been changed. A useful feature of the Dynix client not at present found in GeoPac is its use of a .ini file to specify the EIO viewers. This should allow a wide range of different 'viewers' to be specified for a range of different EIOs and is therefore more flexible and potentially more powerful than the present GeoPac approach.
Although it would be entirely possible to implement a basic
demonstrator project without it, further software development is
required to create a more sophisticated user-friendly
environment for searching a distributed catalogue of the
Internet and retrieving/viewing electronic resources as
envisaged in the full CATRIONA model (See Section 6 and
Appendix B). On the basis of the results of this study there is
good reason to suppose that a demonstrator based on the full
model is feasible. A range of Z39.50 clients have been
examined within the project, and whilst none of these has all of
the facilities required for the full model, most of these facilities
have been seen to work in one client or another. Other
requirements for implementing a demonstrator based on the
full model are also in place, or will be in future. For example,
on the co-operative cataloguing/'Union' catalogue front, it is
now clear that both OCLC and RLG will be creating USMARC
records describing electronic information objects (EIOs), and it
seems probable that co-operative cataloguing of Internet
resources will follow these initiatives; indeed the OCLC project
entails a co-operative element, as do the RLG plans for
cataloguing electronic resources (OCLC, 1994) (Washburn,
1995).
Note: Appendix D provides further information on the Z39.50
OPACs described above and of the USMARC records
retrieved, together with details of the functionality of the Z39.50
clients examined at the time of the study.
___________________________________________________________________________
Chapter 6. Summary of Development
Requirements
Development here is taken in the widest sense and includes
such things as skills development as well as software
development.
N.B. The client and server requirements list at 6.1 below arose from a consideration of Version 3 of the CATRIONA model. These development requirements were utilised in the examination of Z39.50 client functionality and in the exercise to produc e an estimate of costs in the proposed follow-up demonstrator project. The additional requirements listed at 6.2 below arose out of Version 4 of the model. This is described in outline in Section 3 and in detail in Appendix B and is an enhancement of Version 3 of the model (which is not otherwise produced in this report). Appendix B shows how requirements relate to specific aspects of the model.
If a reliable URN to URL mapping service proves feasible (Daniel, 1994b), this would cease to be a problem, although the problem of how URLs are to be reliably kept up to date in a mapping service still has to be tackled. The possibility that a URN to URL resolution service might utilise CATRIONA-type library catalogues as a stable source of URN to URL mappings is worth considering.
(c) In order to inter-work with Z39.50 servers containing IAFA templates (say), other projects need to be identified which utilise Z39.50 and IAFA or which can be persuaded to do so. IAFA templates would have to be registered as Z39.50 record format and clients and servers built to handle them.
(d) Stable national Veronica, Archie and WWW indexing services will be required, possibly at duplicate sites (e.g. BUBL and NISS).
(e) Librarians must have access to information services which enable them to keep up to date with new publications in the electronic world, just as they now have this for the hard-copy world. These would include publishers' catalogues, online and printed, and servic es like BUBL.
In terms of moving the demonstrator beyond the experimental stage and into the embryonic operational system stage, there is a requirement for the installation of a number of Z39.50 catalogues at UK sites. Only one or two such catalogues exist at present, but a large number of libraries are either currently out to tender for new systems or will be within the proposed span of the project and it is almost certain that this requirement will begin to be met within the next six to nine months.
There is also a need to consider whether the proposed FIGIT
(Follett Implementation Group for Information Technology)
subject services should not also be Z39.50 based to enable
them to be integrated fully within a CATRIONA-type distributed
catalogue. Ideally, they should also be MARC based so that
additional client and server development would not be required
to allow existing clients and servers to handle their records.
Failing this, the record-types utilised should be registered for
use with Z39.50 and costs for the development of client and
server functionality for handling these record formats should be
built into the CATRIONA project.
(Note: it is known that the initial FIGIT subject services will be
based on IAFA, WWW and WHOIS++ technology. In which
case integration with CATRIONA may rely upon Z39.50/WWW
and Z39.50/WHOIS++ gateways).
___________________________________________________________________________
Chapter 7. CATRIONA Demonstrator: General
Points
and that, perhaps more crucially:
7.2.1 Because the Z39.50 clients and server programs utilised in the model can and, it is assumed, will come from a range of different suppliers, mainly library system suppliers (Geac, Dynix, SLS etc.), and because any organisation, institution, company or service can become part of the distributed catalogue simply by utilising Z39.50 protocols and the enhanced MARC standard and by letting the world know how to access their catalogue.
7.2.2 Because the model allows other solutions (IAFA template catalogues, Veronica, Archie and so on) to be incorporated and utilised where necessary or appropriate, albeit in the limited ways described in Appendix B, stages 6 and 7. Full integration requires the full integration of the services into the Z39.50 universe (which is proposed here) or special development of Z39.50 client functionality outwith Z39.50 (which is not at this stage proposed here). It is not known to what extent gateways will allow full integration into the Z39.50 universe (e.g. the Z39.50/WHOIS++ gateway alluded to as a possible future development in the ROADS FIGIT proposal) (ROADS, 1995) but the model will also be open to such possibilities.
7.2.3 Because the latest version of the model described in Appendix B allows for any NIDR client, past, present or future to be incorporated and utilised in the network navigation/resource discovery process
Any demonstrator project and, indeed, any working system based on the model would therefore necessarily be 'open' also. Other libraries, organisations, services and suppliers could join at any time, simply by adopting the same, or a similar, model, and it is by this means that a more comprehensive catalogue of Internet resources might eventually emerge.
___________________________________________________________________________
Chapter 8. CATRIONA Demonstrator: Specific Phase II
Proposals
The aim would be to move, over a three year period, towards an embryonic but stable operational (i.e. non-experimental) system based upon the CATRIONA model or some adaptation of it - i.e. a system and associated infrastructure capable of supporting a sustainable, distributed and extendible approach to networked resource discovery, description, retrieval, and presentation, based upon existing library procedures, practices, and standards (e.g. Z39.50, MARC) and incorporating library system based access control mechanisms.
It is not envisaged, however, that it would be possible to move directly to the installation of the fully operational system immediately. Rather, a three year programme is envisaged, as indicated below.
8.2.1. Identification of a representative mix of network resources for cataloguing, utilising the BUBL subject tree, local resource identification at Scottish libraries, liaison with FIGIT projects creating new resources. (Including, in time, the Stirling On-Demand Publishing project). If implemented at BUBL, items from the Alex catalogue might also play a part.
8.2.2. Subject-area based division of resource discovery duties between Scottish libraries and other external groups (e.g. SOSIG, OMNI) based on the BUBL/NISS LIS-subjects approach, in order to help identify network resources for cataloguing and build an infrastructure for future co-operative activity in this area. Liaison, co-operation and, if possible, shared cataloguing agreements with organisations such as OCLC and RLG.
8.2.3. The storage of e-resources identified for cataloguing and not already located either at the site of a participating library or on BUBL on a centrally located experimental server.
8.2.4. Creation of network-accessible MARC-based catalogue of network resources so identified on the same centrally located experimental server. This to be Z39.50 compliant but also to allow, via a link to BUBL, more basic x.29 and Telnet access to support the needs of the widest possible range of HE institutions (Telnet from BUBL to a VT session on supplier Z39.50 server, for example). Access via WWW would also be investigated.
8.2.5. Development of skills and experience in project assistants and transfer of these to staff in participating libraries (cataloguing, systems, reference) by means of on-site training programmes.
8.2.6. Working in conjunction with a range of library, library systems supplier, library bibliographic utility organisation partnerships, establish technical and other requirements for developing facilities and skills at individual participating libraries to the level required for active participation in the proposed distributed system in years 2 or 3 (or earlier if this is possible).
8.2.7 Further investigation at two or more sites of existing client
capabilities and software development requirements with non-
parallel distributed searching being utilised and examined,
although at this stage only one central CATRIONA Z39.50
OPAC would exist, other sites will exist elsewhere (Brigham
Young and perhaps OCLC and RLG) and some Scottish
universities (e.g. Strathclyde) will be implementing Z39.50
OPACs.
In addition, there may be FIGIT subject services utilising
Z39.50 and MARC if the recommendations of this report are
heeded.
8.2.8 If appropriate, initial development of client functionality in the areas described in Section 6 of the report (not yet prioritised). However, development requirements may have to be more clearly defined in Year 1 in consultation with suppliers.
8.3.2 As far as is appropriate, distribute centrally-held resources and associated enhanced records to local sites from central server (essential to model that sites only hold records of items stored locally, but a reliable mechanism for informing sites of URL changes at BUBL (or indeed other JISC services such as NISS) might be a workable alternative, in which case only catalogue records would be distributed in many cases.
8.3.3 Contribute all centrally held records to various participating co-operative cataloguing utilities to establish the basis of a working system based on the model.
8.3.4 Train local staff to participate in the use and further development and growth of the system.
8.3.5 Turn what remains of the BUBL-based central catalogue and document server into a subject service covering LIS and associated areas such as general reference tools (Internet-wide directories, electronic glossaries, dictionaries etc.)
8.3.6 Identify Z39.50 OPACs outwith the project which might usefully be encompassed within the distributed catalogue.
8.4.2. Enhancement of existing local and 'Union' catalogues to incorporate descriptions of existing and FIGIT-created network resources and data to enable their retrieval and display on campus and library workstations.
8.4.3. Refinement and/or enhancement of existing library system Z39.50 OPAC clients to enable the integration of the retrieval of traditional resources with that of the retrieval of electronic resources.
8.4.4. An infrastructure for the creation and maintenance of a distributed OPAC system for electronic resources in the UK and, ultimately, the wider world.
8.4.5. Increased skills and awareness in library staff of the requirements of networked resource discovery, description and retrieval
8.4.6. Improved service to the communities served by UK academic libraries.
8.4.7. Improved understanding of management issues relating to co-operative approaches to network resource access
8.4.8. Help to establish a world-wide UK library influence in areas relating to cataloguing networked electronic resources and finding, retrieving and displaying networked electronic resources.
8.4.9. Advise JISC on related requirements with regard to the network navigation support efforts required from centrally- funded JISC services (e.g. stable and duplicated Veronicas, Archies and similar WWW indexing services)
___________________________________________________________________________
Chapter 9. CATRIONA Phase II Requirements: Skills,
Personnel, Organisation
Factors likely to affect the cost of software development:
To a lesser extent, there is a degree of uncertainty about some of the other costs involved, with the cost of travel and peripheral equipment likely to vary according to the exact shape and form agreed for the follow-up project and the position of funding bodies and libraries on funding relating to peripheral equipment such as microcomputers.
The estimates detailed below are therefore only a rough guide to the likely costs of a demonstrator project. They assume company consultancy fees in Year 1 and software development costs in years 2 and 3.
___________________________________________________________________________
http://www.bubl.bath.ac.uk/BUBL/maincatriona.html
gopher://bubl.bath.ac.uk:7070/11/Link/Catriona
http://www.bubl.bath.ac.uk/BUBL/Tree.html
gopher://www.bubl.bath.ac.uk:7070/11/Link/Tree
Chapter 10. CATRIONA and the BUBL Subject Tree
Project:
Further information on the CATRIONA feasibility study, on
other related information and projects, and on the BUBL
Subject Tree will be found on the BUBL Information Service at
URLs:
Further Information
This information will be added to and enhanced by the BUBL Information Service as information vital to the library community and its users regardless of whether or not the CATRIONA project continues into Phase II and beyond.
More information on the Subject Tree project and its relationship to the CATRIONA project is provided in Appendix I.
___________________________________________________________________________
References
ANIR (1994). The Working Group on Access to Networked
Information Resources of the Information Services Sub-
Committee of the UK Higher Education Funding Councils.
Report.
ANSI/NISO (1994). ANSI/NISO Z39.50-1994, Information retrieval application service definition and protocol specification. Available from ftp://ftp.loc.gov/pub/z3950
Bowman C M et al (1994). Harvest: a scalable, customisable discovery and access system. Technical Report CU-CS-732- 94, Department of Computer Science, University of Colorado, Boulder. Available from ftp://ftp.cs.colorado.edu/pub/cs/techreports/schwartz/Harvest.F ullTR.ps.Z
Christian E (1994) The Government Information Locator Service (GILS): report to the Information Infrastructure Task Force. May 2, 1994. Available from ftp://info.er.usgs.gov/public/gils/gils.txt
Daniel R Jnr (1994). URC scenarios and requirements. Internet draft. November 21, 1994 (Work in progress). Available from ftp://nic.nordu.net/internet-drafts/draft-ietf-uri- urc-req-00.txt
December J (1994) New spiders roam the Web. Computer- Mediated Communication Magazine 1: 3. Available from http://www.rpi.edu/~decemj/cmc/mag/1994/sep/spiders.html
Dempsey L (1994) Network resource discovery: a European library perspective. In Libraries, networks and Europe: a European networking study. N Smith (ed) London: British Library Research and Development Department. Available from gopher://ukoln.bath.ac.uk/00/Publications/ResDes/europe.rtf
Deutsch P et al (1994). Publishing information on the Internet with anonymous FTP. Internet draft. September 1994 (Work in progress). Available from ftp://nic.nordu.net/internet-drafts/draft-ietf-iiir- publishing-02.txt
Dillon, M et al (1993). Assessing information on the Internet: toward providing library services for computer mediated communication. Dublin, Ohio: OCLC.
Gaynor E (1994). Cataloging electronic texts: the University of Virginia Library experience. Library Resources and Technical Services 38: 403-413.
Giordano R (1994a). Notes on operations: the documentation of electronic texts using Text Encoding Initiative headers: an introduction. Library Resources and Technical Services 38: 389-401.
Giordano R (1994b). URLs, etc. TEI discussion list. 29 December, 1994. Available email: TEI-L@UICVM.EARN.
Herr-Hoyman D (1994). Re: Library standards and URIs. URI discussion list. 29 December, 1994. Available email: URI@BUNYIP.COM.
Koster M (1994). Re: URCs and IAFA templates. URI discussion list. 13 October, 1994. Available email: URI@BUNYIP.COM.
Hazelden L (1994). SilverPlatter Information Ltd. Personal communication, 19 December, 1994.
MARBI (1991a). Discussion paper no. 49: dictionary of data elements for online information resources. Library of Congress. May 1, 1991. Revised July 15, 1991.
MARBI (1991b). Discussion paper no. 54: providing access to online information resources. Library of Congress. November 22, 1991. Revised April 13. 1992.
MARBI (1993a). Proposal no. 93-4: changes to the USMARC bibliographic format (computer files) to accommodate online information resources. Library of Congress and OCLC Internet Resources Project. November 20, 1992. Revised March 29, 1993. Available from gopher://marvel.loc.gov:701/00/.listarch/usmarc/93-4.doc
MARBI (1993b). Discussion paper no. 69: accommodating online systems and services in USMARC. Library of Congress and OCLC Internet Resources Project. April, 30 1993. Revised August 4, 1993.
MARBI (1993c). Proposal No. 94-2: addition of subfields $g and $3 to field 856 (electronic location and access) in the USMARC holdings/bibliographic formats. Library of Congress. December 6, 1993. Available from gopher://marvel.loc.gov:701/00/.listarch/usmarc/94-2.doc
MARBI (1993d). Proposal no. 94-3: addition of subfield $u (Uniform Resource Locator) to field 856 in the USMARC Bibliographic/Holdings Formats. Library of Congress. December 6, 1993. Available from gopher://marvel.loc.gov:701/00/.listarch/usmarc/94-3.doc
MARBI (1994a). Proposal no. 94-9: Changes to the USMARC Bibliographic Format to Accommodate Online Systems and Services. Library of Congress and OCLC Internet Resources Project. May 6, 1994. Revised July 20, 1994. Available from gopher://marvel.loc.gov:701/00/.listarch/usmarc/94-9.doc
MARBI (1994b). Proposal no. 95-1: Changes to Field 856 (Electronic Location and Access) in the USMARC Bibliographic Format. Library of Congress and Federal Geographic Data Committee. December 2, 1994. Available from gopher://marvel.loc.gov:701/00/.listarch/usmarc/95-1.doc
McBryan O (1994) GENVYL and WWWW: tools for taming the Web. In: First International World-Wide Web Conference. Advance proceedings. May 25-27. Geneva.
Mealling M (1994). Encoding and use of Uniform Resource Characteristics. Internet draft. July 8, 1994 (Work in progress) Available from http://www.acl.lanl.gov/URI/archive/uri- 94q3.messages/21.html
Mitra, Weider C & Mealling M, (1994) Uniform Resource Names. Internet draft. October 20, 1994. (Work in progress). Available from ftp://nic.nordu.net/internet-drafts/draft-ietf-uri- resource-names-03.txt
Neuss, C and Hofling, S (1994) Lost in hyperspace? Free text searches in the Web. In: First International World-Wide Web Conference. Advance proceedings. May 25-27. Geneva.
OCLC (1994). Building a Catalog of Internet Resources project. Available from http://www.oclc.org/oclc/man/catproj/catcall.html
ROADS (1995) Private document supplied by Lorcan Dempsey on behalf of the ROADS consortium, 9 March, 1995.
SilverPlatter Information (1994). An Introduction to SilverPlatter's ERL technology.
Sperberg-McQueen C M and Burnard L eds.(1994) TEI P3: Guidelines for electronic text encoding and interchange. Oxford and Chicago: The Text Encoding Initiative. Available from http://etext.virginia.edu/TEI.html
van der Werf T (1994). InfoServices: cooperation between the national research network service and the National Library in the Netherlands. Journal of Information Networking 2: 13-22.
Washburn B (1995). Research Libraries Group. Personal communication, 10 January, 1995.
Weibel S (1994). OCLC. Outline notes for address to IETF Meeting in San Jose, December 1994. Personal communication and subsequently posted on URI list, 12 December 1994. Available email: URI@BUNYIP.COM.
Weider C & Deutsch P (1994) A vision of an integrated internet information service. Internet draft. July 31, 1994. (Work in progress). Available by email from ds.internic.net Send the following message: document-by-name RFC1727
Weider C (1994) The Internet anonymous FTP archive templates: towards an Internet resource location system. Journal of Information Networking 1: 256-260.
Weider C (1995) Bunyip Information Systems Inc. Personal communication, 3 February 1995.
VINE (1994) VINE:Theme issue: Z39.50 and SR. 97.
___________________________________________________________________________
___________________________________________________________________________
It is worth reiterating that whilst it has been possible to refine
and improve the model as a result of research carried out
under the project, it is assumed that further adjustments will be
necessary in the light of applying the model in a real working
situation in a demonstrator project.
The model description is built up in stages by following the
options that would be available to a user of a CATRIONA-type
OPAC client. There are 14 stages in this illustrative description,
each of which is based upon different kinds of circumstances
which the user might face:
Stages 1-8 address a situation in which decisions as to when
and how to escalate a search beyond the local catalogue are
taken by the user. Stages 9-10 address the situation in which
these 'decisions' are controlled or influenced by the system
itself. It was felt that the best way of describing the model was
to present the user-controlled escalation options first. This
should not be taken to imply either that this is the first choice
that would be 'offered' to the user or that there is any reason to
believe that this would normally be the user's preferred choice.
It may well be that most users would prefer to let the system
handle such 'decisions'. Stages 11-14 cover some additional
situations which are considered pertinent.
What is envisaged, then, is that the CATRIONA-type OPAC
client (which would come in many different supplier-specific
forms) would present the user with a number of the usual
search options and also an option to 'search the whole Internet'
or something of that kind. In steps 1-8 of the description of the
model the assumption is that the user has chosen this latter
option. As a result she is presented with a list of individual
cataloguing and indexing services, one of which is the local
OPAC, but most of which are remote. She may choose to
search each individually or to mark some or all for parallel or
distributed searching.
The services will fall roughly into the categories listed at 1-8
above. Whether or not they will be categorised in this way for
the user has yet to be decided. In the illustration, the user
escalates the search gradually from 1 to 8. However, she will
not necessarily proceed in this way in a real situation. She may
well jump to step 3 or 7 on the basis either of her own
experience of her current type of problem or on the basis of
information provided by the system help.
The first record describes a hard copy book. She notes the
classmark in the usual way, aiming to go to the shelf and
retrieve it later.
The second record describes an electronic version of an article
by a lecturer at her institution. The MARC record she has
retrieved has a URL in the 856 field (subfield $u) and the
CATRIONA client responds to this automatically by placing a
'link' button on the screen. The user points to the 'link' button
and clicks the mouse button. The CATRIONA client responds
by automatically loading a WWW client such as Windows
Mosaic and feeding it the URL for the article. Mosaic locates
the article on a local server and transfers it to the user's PC,
loading an appropriate viewer to display it if necessary.
The third record describes the BIDS Embase service (the user
used a very general subject heading). The user indicates a wish
to search this service, at which point the system asks her to
identify herself by her barcode or PIN (Personal Identification
Number). If she has access rights, the CATRIONA client loads
a communications package from the client PC and makes a
Telnet connection to the Embase service allowing the user to
perform a new search on Embase.
___________________________________________________________________________
The user would 'mark' the catalogues then enter a request
which the CATRIONA OPAC client would send to all of the
Z39.50 OPACs in parallel. The OPACs would return the results
showing the number of hits. The user might then choose to
examine results from individual servers. Ideally these will be
returned sorted for ease of use (this requires Z39.50 Version 3).
Another option that might be available to the user is the option
to combine some or all of the result sets into a single sorted set
with duplicates filed together. Depending upon the hardware
limitations this option may only be available if the result sets
fall within certain size limits. Eventually the user would choose
a number of records that describe resources relevant to her
enquiry. One of these may be a hardcopy resource which in
some cases she may be able to 'reserve' or order through Inter-
Library Loan. Another may be an electronic resource which will
again be located at a URL specified in the MARC record (or,
perhaps, ultimately at a URL or URLs specified via a URN).
Once again, the 'link' button will appear and if the user clicks on
it a suitable client/viewer combination will be loaded by the
OPAC client and the resource will be retrieved from the
appropriate site and displayed for the user.
EIOs catalogued on these OPACs will usually be held in many
different places on the Internet (they will be held at the sites
which have catalogued the EIO themselves). However, the EIO
retrieval mechanism will be the same as at B.2 above. The
user will choose a retrieval record which will contain at least
one, but possibly several, URLs (or, perhaps, ultimately, a
single URN which will 'point at' several URLs through a
resolution service). If there is one, the client will load an
appropriate NIDR client (such as Mosaic or Cello), pass it the
URL or the URN, and the EIO will be retrieved and
displayed/played. If there are several, a user choice will have to
be made. Hopefully, this will be an informed choice based on
system-provided information. Ultimately, it may be that the
system will make the choice intelligently for the user on the
basis of local information on network conditions - i.e. it will
choose which URL to try first, which to try next and so on. The
exact process utilised will depend upon the extent to which a
practical URN to URL resolution service is ultimately
developed.
(b) Updating URLs
The method for updating URLs is not yet fully established. An
automatic solution may only develop as these catalogues are
established. There are already robot-type programs which can
check for broken links and some automatic mechanism for
updating the links may develop. In the CATRIONA model at
present each site would have to send in an updated record to
the Union catalogue when the local URL changes.
If a reliable URN to URL mapping service proves feasible this
would cease to be a problem, although the problem of how the
URLs are to be kept up to date in the mapping service would
still have to be tackled.
The assumption behind the CATRIONA project is that the
Internet will be catalogued in a distributed fashion. In order to
carry out a search of this whole distributed catalogue, therefore,
it is necessary for the OPAC client being utilised to be able to
locate and search all of the individual OPACs that make up the
distributed catalogue. In existing clients, this information is
programmed into the client by the system administrator.
However, the assumption in the CATRIONA model is that this
information will initially be located on the local OPAC which will
contain catalogue records for other OPACs selected by the
library and uploaded to the client as required. It is further
assumed that any given local OPAC will only contain a small
subset of all of the available OPACs and that much of the
information about the rest of the distributed catalogue of the
Internet will come from other OPACs and, in particular, the
Union catalogues. It is likely that Internet OPACs will be
categorised according to their subject strengths and that users
will be able to search and locate their catalogues for records of
other OPACs strong on a particular subject category. The user
will then be able to conduct a distributed search of a selection
of these OPACs in a similar fashion to that described at B.1 to
B.3 above.
It is assumed that for some time to come, and perhaps
indefinitely, valuable OPACs will exist on the Internet which do
not utilise MARC and that these will also be catalogued in local
and Union OPACs and potentially able to be part of any
distributed search conducted by the user. It is unlikely that the
user will know or care whether the records she retrieves are
MARC records or IAFA templates or TEI headers or GRS or
SUTRS or whatever and there may be nothing in the list of
'other services' the user is presented with to distinguish these
catalogues from some of the others she will be presented with.
They may be categorised by subject area but it is unlikely that
they will be categorised as non-MARC. They will simply be one
of a number of OPACs that the user may or may not choose to
include in her distributed search. When records of Internet
resources are retrieved from these OPACs the client will call up
Mosaic (or whatever) and utilise the URL in a similar way to
that described earlier, although this may require client
development in order that the client should be able to identify a
URL that is not in a MARC 856$u subfield and client and server
development to handle the record formats used. The record
formats used (e.g. IAFA templates) would also have to be
registered as accepted Z39.50 record formats.
If the user chooses to search one of these catalogues, the
CATRIONA client will automatically load a Telnet program or
Web or WAIS client depending on how the non-Z39.50 service
has been described in the local OPAC (i.e. what is in the 856$u
subfield). The user will be logged on to the remote site and will
conduct a search using whatever interface it provides (the
ability to save, edit and resubmit search strategies may be of
value here. In cases where the search client needed to search
the remote catalogue is incapable of using a URL to retrieve a
resource, it will be possible to copy retrieved records containing
URLs to a file on the user's workstation and to convert them to
HTML so that the user can then use Mosaic to retrieve and
display the EIOs found.
The assumption here is that a large number of catalogues at
present and for the foreseeable future will not only not be
MARC, they will not be Z39.50 either. They will be offered to
the user as important sources of Internet metadata but it will
probably not be possible either to include them in any
distributed search of the Internet or to utilise the same search
strategies in them as were applicable to Z39.50 OPACs.
ERL is media independent so that it can accommodate
databases in either CD-ROM or hard disk format. With ERL the
databases can be copied to hard disk and a hybrid CD-
ROM/hard disk solution can be created.
This would operate in a way similar to the other Z39.50
searches described above as the gateway maps the DXP
protocol to Z39.50.
If the user chooses 'search selectively' the system loads details
of all of the catalogues catalogued in the local OPAC into 'other
searches' and allows the user to mark those she wishes to
search, then proceeds as above.
___________________________________________________________________________
___________________________________________________________________________
http://www.bubl.bath.ac.uk/BUBL/catriona.html
to a model described in a series of 26 points at URL:
gopher://ukoln.bath.ac.uk:7070/0/Link/Catriona/Catriona_Model
_Clarification
which was presented over the CATRIONA discussion list to
systems suppliers in September 1994 for feedback. The model
was then revised to produce Version 3 as more information was
gained on the issues involved. Version 4 of the model (See
Appendix B) was developed late in the project. Version 4 is an
enhanced and extended version of Version 3. Version 3 was
used to identify the initial development requirements utilised in
the costing for Phase II (See also Section 6 of the main report).
URNs, URLs, URCs
___________________________________________________________________________
Note: This appendix should be read as an addendum to Section
5 of the report.
Connection details provided:
Butler University
North Carolina State University
The connection information above was used to add the two
databases to the existing range of databases available in
GeoPac as originally installed. Brigham Young University also
provided details of the attribute set (e.g. 1 = Personal Name, 2
= Corporate Name etc.) supported by the server and this was
set up in the client so that the client search was conducted
under the appropriate parameters. Details of the records were
also provided. These records had been enhanced to include
URLs in 856$u (as recommended by MARBI in Proposal 93-4).
On retrieval of the records the client's 'link' button was
highlighted and clicking on this had the effect of making the
OPAC client load a suitable 'viewer', and pass it the URL so
that it could locate, retrieve and display the electronic
information object (EIO) described by the MARC record.
The viewer loaded was the WWW client Mosaic and the EIO
was a full text article on a gopher located at Minneapolis:
The Internet Gopher Protocol; a distributed document search
and retrieval protocol: Alberti, Bob
Marc record type USmarc
008 940628s1992bbbbmnubbbbbbrbbbb00000bengb
BYU Record 2
The viewer loaded was a .gif image viewer and the EIO was a
.gif file located at Brigham Young University:
Photograph of Joseph F. Smith: Johnson, Charles Ellis
Marc record type USmarc
Butler Record 1
Mainstreaming our online catalogs: Kambitsch,
Timothy G.
ALA Washington Office newsline ALAWON : an
electronic publication of the
American Library Association Washington Office.
5. Number of hits given before records are received
Appendix B
During the course of the project, a number of attempts have
been made to describe different incarnations of the CATRIONA
model for project participants and others. Judging by the
evidence of later discussions of the description with some of
the participants, none of these attempts has been entirely
successful. It is believed that this has been due partly to an
attempt to describe all of the model all at once rather than in
piecemeal fashion, and partly to a failure to put the description
in a real context. What is presented here, therefore, in Version
4, is an attempt to build a complete picture of the model
gradually, and to do so in the context of a real user, in this case
a student, utilising a system based on the model. (Version 3 of
the model description was also based on this approach).
CATRIONA Model Version 4:
Illustrative Description of User Interaction
STAGES 1-8: USER-CONTROLLED ESCALATION OF
SEARCH
STAGES 9-10: SYSTEM-INFLUENCED ESCALATION
STAGES 11-14: OTHER POSSIBILITIES
___________________________________________________________________________
B.1. STAGE 1: LOCAL OPAC SEARCH
In this illustration, the user chooses to perform a subject search
in only the local OPAC. She performs her search in the usual
way and retrieves a number of hits, three of which are relevant
to her information problem.
Stage 1 Identified Development Requirements
Client / Server Requirements
(1) PC client (plus, ideally, Mac and X Windows and
VTxxx).
(13) Ability to call up one of several NIDR clients as
appropriate (e.g. WWW. Gopher, WAIS) (use .ini file as in
Mosaic).
(a) Ability to offer Telnet access to service if and only if
user permissions permit.
(b) Ability to auto logon to remote service using username
and passwords.
(c) Ability to suppress display of username and password
stored in (MARC) record except at system administrator permissions
level.
(d) Ability to suppress retrieval/display of records of
commercial services when user is not local.
Cataloguing Requirements
(a) Ability to catalogue any and every possible type of
Internet resource and service.
(b) Ability to record username and password of
commercial services in suppressible (MARC) field.
(c) Ability to record that this is a special catalogue record
only to be retrieved/displayed if the search is a local one.
URL Requirements
(a) URLs for local electronic resources and remote major
services must be kept up to date. It is assumed
that this will be handled as described in Section 3.2 of
the main report.
___________________________________________________________________________
B.2. STAGE 2: DISTRIBUTED OR PARALLEL SEARCH OF OTHER 'ORDINARY'
OPACS AND/OR CATALOGUES OF SUBJECT- SPECIFIC RESOURCES
Having exhausted the possibilities of the local OPAC, the user
chooses to perform a distributed search of a number of remote
Z39.50 OPACS catalogued in the local OPAC and offered in
response to the user clicking the "search other catalogues"
button. In the proposed CATRIONA Phase II project, these
would include a number of Scottish Z39.50 OPACs with subject
specialisations, together with any available Z39.50 catalogues
of subject-specific services. The local library might have a
special arrangement with other libraries in a given geographical
area (e.g. an agreement to share the work of Internet resource
discovery and cataloguing on a subject basis or to purchase or
create electronic resources co-operatively on a subject basis)
and the CATRIONA client would offer a choice based on such
considerations.
___________________________________________________________________________
Stage 2 Identified Development Requirements
Client / Server Requirements
(2) Data for accessing other Z39.50 OPACs not on menu
can be uploaded from catalogue records.
(3) Distributed parallel searching.
(4) Group of databases can be selected by user for parallel
searching.
(5) Number of hits given before records are received.
(6) Numbers of hits from several servers are given
separately.
(7) Warning if result sets are too large.
(8) Stop button (does this need to be individual to a server
session?)
(9) Option to merge results from several servers.
(16) Servers to appear in catalogue.
(10) Option to sort results (URNs or pseudo URNs?)
(11) Filing duplicates together (URNs or pseudo URNs?)
(e) Ability to display subject category of catalogued remote
OPACs.
Cataloguing Requirements
(d) Records must be in a form that can be searched and
retrieved using standard protocols such as Z39.50.
(e) Ability to catalogue and subject index other Z39.50
OPACs and record Z39.50 information needed to
set up client-server dialogues.
(f) Ability to store several alternative URLs in catalogue
record.
URL Requirements
(b) URLs for remote electronic resources must be kept up
to date. It is assumed that this will be handled as
described in Section 3.1.2.3 of the main report.
Progress towards creating a URN to URL resolution
service will be relevant here.
URN Requirements
(a) Filing duplicates together may require URNs or
pseudo-URNs.
___________________________________________________________________________
B.3. STAGE 3: OPTION 2 PLUS UNION CATALOGUES
This search is similar to the previous search (See B.2), but in
this case one or more Union catalogues (e.g. OCLC, RLG,
CURL) are also marked for searching, having been loaded with
the other Z39.50 OPACs catalogued locally in response to the
user choosing the "search other catalogues" option. Perhaps
the user will choose to limit her search to EIOs only. Note,
however, this is just an option, she may, equally, decide not to
limit it.
___________________________________________________________________________
Stage 3 Identified Development Requirements
Client / Server Requirements
(18) Ability to do a Z39.50 OPAC EIO/Internet search only
(i.e. to search only for electronic resources instead
of hard-copy resources as well).
(f) Ability to handle multiple URLs.
(g) Ability to make an intelligent choice between
alternative URLs.
Other Requirements
(a) Union Internet catalogues . The existence
of 'Union' Internet catalogues is a requirement here. OCLC are currently
compiling a catalogue of Internet resources in their project 'Building a
Catalog of Internet Resources'. This catalogue will be made publicly
available. RLG are planning to add Internet resources to their Z39.50
Zephyr catalogue at the beginning of 1995 (Washburn, 1995). Similar
records may be available on the CURL database in future. Given that such
co-operative 'Union' catalogues will exist, it is reasonable to assume
that they will be utilised as a more efficient way of searching the whole
Internet than searching every single local OPAC.
URL Requirements
(See above - Updating URLs in Union Catalogues)
___________________________________________________________________________
B.4. STAGE 4: SEARCH OF OTHER (MARC) Z39.50 OPACS SERVED UP BY UNION
AND OTHER CATALOGUES
In this case the user conducts a distributed search as in B.2
and B.3 above but either gets no hits, or no useful hits, or
insufficient hits for her purposes. At this point she might have
the option of searching for records of additional Internet OPACs
which might, in turn, satisfy her enquiry. It is assumed that her
search would be subject-based and that the catalogue records
describing other Internet OPACs would also have general
subject classifications. The result of the user's search would be
a new list of "other Internet catalogues" being offered by the
CATRIONA client for her to mark for distributed searching. The
search, retrieve and display/view process would then proceed
as described in B.2 and B.3. In some cases the document or
resource found in one or other of the catalogues might be
charged for and a means of handling this would be built into the
client-server dialogue.
___________________________________________________________________________
Stage 4 Identified Development Requirements
Client / Server Requirements
(15) Access control and charging mechanisms.
(h) Ability to conduct a subject search for records of other
Z39.50 OPACs
___________________________________________________________________________
B.5. STAGE 5: SEARCH OF Z39.50 OPACS CONTAINING
NON-MARC RECORDS (e.g. IAFA TEMPLATES AND TEI HEADERS)
At some point in her distributed parallel searching of other
Internet catalogues, it is envisaged that the user might
encounter Z39.50 OPACs which are not MARC-based. There
may, for example, be library OPACs which do not use MARC
or other non-library services using IAFA templates, TEI
headers, Z39.50 SUTRS or GRS.
___________________________________________________________________________
Stage 5 Identified Development Requirements
Client / Server Requirements
(19) Ability to handle URLs in SUTRS/GRS i.e. not in
MARC 856 (client and server requirement).
Cataloguing Requirements
(g) Ability to record Z39.50 information needed to set up
client-server dialogues with non-MARC Z39.50 OPACs.
(h) Records must be in a form that is compatible with
developing Internet standards.
Other Requirements
(c) In order to inter-work with Z39.50 servers containing
IAFA templates (say), other projects need to be identified
which utilise Z39.50 and IAFA or which can be
persuaded to do so. IAFA templates would have to be
registered as Z39.50 record format and clients and servers
built to handle them.
___________________________________________________________________________
B.6. STAGE 6: SEARCH OF NON- Z39.50 CATALOGUES
Some of the "other sources" presented to the user as options
for a remote search of other Internet catalogues will be
presented under a separate heading which will indicate that
these are usually to be utilised after other options have been
exhausted and also that they are non-Z39.50 and must be
searched individually using whatever search language they
employ.
__________________________________________________________________________
Stage 6 Identified Development Requirements
Client / Server Requirements
(14) Save search strategy.
(17) Ability to conduct distributed (server by server) search
for other non-Z39.50 Internet OPACs.
(20) Close down NIDR client session button in OPAC client.
(i) Ability to copy retrieved records containing URLs from
a remote non-Z39.50 OPAC to a user file.
(j) Ability to convert document containing URLs to HTML
in order to allow retrieval of EIOs by NIDR client.
__________________________________________________________________________
B.7. STAGE 7: SEARCH INTERNET INDEXING SERVICES
Some of the kinds of 'other sources' dealt with in B.6 will not be
non-Z39.50 OPACs but Internet indexing services such as
Veronica or Archie or a WAIS site. Acting on the information
found on the local catalogue record regarding the service
concerned, the CATRIONA client would load the appropriate
client for the service concerned, make the connection, and
allow the user to conduct a search which would usually result in
the eventual retrieval and display of an appropriate EIO. In the
case of Archie there may be a facility to capture screens and
HTML-ise the results to that the WWW client can start an
actual FTP (File Transfer Protocol - see Glossary) session.
The possibility of utilising Internet indexes such as Veronica
and Archie is likely to continue to play a part in any Internet-
wide search for the foreseeable future. These will be options on
the menu of other services and, once again, an appropriate
client will be called up to enable access. As with category 6,
they will be accessed individually and will require a service-
specific search strategy.
___________________________________________________________________________
Stage 7 Identified Development Requirements
Client / Server Requirements
(k) Ability to load appropriate client to conduct search of
Internet indexing services such as Archie or
Veronica.
Cataloguing Requirements
(i) Ability to catalogue Internet indexing services such as
Archie, Veronica etc. (A specific example of catalo
guing requirement 1).
Other Requirements
(d) Stable national Veronica, Archie and WWW indexing
services will be required possibly at duplicate sites
(e.g. BUBL and NISS).
___________________________________________________________________________
B.8. STAGE 8: SEARCH NETWORKED CD-ROMS OR OTHER
DATABASES
The user may retrieve records locally describing locally held or
remote CD-ROM titles accessible over the network from her
workstation. Ideally the client would seamlessly connect the
user to the preferred service and allow her to use it.
___________________________________________________________________________
Stage 8 Identified Development Requirements
Client / Server Requirements
(25) Gateway to search ERL-compliant databases.
(26) Ability to load networked CD-ROM titles and to handle
associated memory requirements.
Cataloguing Requirements
(j) Ability to catalogue CD-ROM titles and record access
information for them. (a specific example of
cataloguing requirement 1).
Note: ERL
SilverPlatter is using ERL (Electronic Reference Library) as the
means SilverPlatter to provide Wide Area Network (WAN)
access to large databases using TCP/IP. ERL is made up of
two software components: the ERL clients (retrieval interface)
and the ERL server (search engine). ERL runs over DXP (Data
eXchange Protocol) which was designed by SilverPlatter to
facilitate client/server interaction while keeping the functionality
of stand-alone retrieval systems. Thus the DXP specification
includes Boolean searching, thesauri and hot links as well as
multi-database searching. Although several databases
available on one server can be searched, at present the
databases are not searched in parallel and the results are not
sorted. This development is planned. In future DXP will also
specify meta-thesauri for multiple database searching. Secure
access to the databases is also provided as only authorised
users can log in.
Z39.50
In ERL, the DXP co-exists with various standards including
Z39.50 for libraries. Library clients written to Z39.50 can access
all ERL compliant databases through a gateway. Ameritech is
developing a gateway to allow their retrieval clients (WinPAC,
ProPAC, TermPAC and MacPAC) to search ERL-compliant
databases from the same Z39.50 client interface and
SilverPlatter has programs to share technology with vendors
who want to develop gateways from their clients to ERL
compliant databases.
Limiting and paying for remote access
Remote access can be arranged and controlled as ERL
administrators control the security for multiple server access
over the Internet. 'Servers are configured to indicate the
machines and/or users who have permission to use one or
more of the databases located on that server, through either
user names and passwords or TCP/IP address ranges.
Authorised users are presented with a single list of available
databases, which may reside on a number of ERL servers'
(SilverPlatter Information, 1994).
Payment for remote access would be arranged with the
individual information providers through SilverPlatter
(Hazelden, 1994).
___________________________________________________________________________
STAGES 9-10: SYSTEM-INFLUENCED ESCALATION
This section of the description of the model deals with the
possibility of system-influenced escalation of searches beyond
the local site as a result of failed local OPAC searches.
B.9. STAGE 9: FAILED KNOWN-ITEM SEARCH
This part of the illustration deals with a situation where the user
conducts a known item search of the local OPAC. The search
fails. In addition to presenting the user with 'near hits' from the
local OPAC, the system also suggests that the user try a
search of other Internet OPACs offering the user a choice of
'search none', search all', 'search selectively' (user choice). In
the event of the user choosing 'search all', the system sends
the search to each of the Z39.50 catalogues it 'knows' about. If
it finds a hit it informs the user and offers to stop at that point. If
all Z39.50 searches fail, it offers a user-driven option to search
the other catalogues and indexes it 'knows' about (of course, if
the user knows the URN, any URN to URL resolution service
would be brought into play here).
Stage 9 Identified Development Requirements
Client / Server Requirements
(23) Ability to respond to a failed known item search by
offering to search other Internet resources
(24) Ability to automatically search other OPACs in turn for
known item and offer to stop searching when hit is
found.
___________________________________________________________________________
B.10. STAGE 10: SYSTEM RESPONSE TO SUBJECT SEARCH
This part of the illustration deals with a situation where the user
performs a subject search in the local catalogue. Since it is
possible either that a subject search will produce no hits or that
it will produce hits which are either inappropriate or insufficient
for the user's purposes, the system would always respond by
offering to search other Internet sources. In the longer term, it
might be possible that systems will respond intelligently to this
situation, utilising the original search terms, together with a
thesaurus, to automatically search for and identify appropriate
subject catalogues on the Internet for the user to search (e.g. a
search for 'quarks' might throw up Physics-strong OPACs as
other sources). In the shorter term, a more likely approach
would be to offer the user the chance to choose from a list of
subject catalogues, those most appropriate to her search.
___________________________________________________________________________
Stage 10 Identified Development Requirements
Client / Server Requirements
(22) Ability to offer 'other subject catalogues' option after a
subject search.
(l) Ability of server to use a thesaurus to identify general
subject terms appropriate to more specific terms
entered by user, and to use the general terms to identify
appropriate subject catalogues to offer the user as a
means of expanding the search.
___________________________________________________________________________
STAGES 11-14: OTHER POSSIBILITIES
This section of the description of the model deals with the
possibility of system-influenced escalation of searches beyond
the local site as a result of failed local OPAC searches.
B.11. STAGE 11: NEW EDITIONS
Having found and scanned two electronic documents via the
local OPAC, the user wants to find out if there are new editions
of the document available elsewhere on the Internet. There
would be an option to do a 'find others like this but more up to
date' option. Having found a later edition of one of the works,
the user asks the librarian whether he can find out if there is a
later edition of the other. After some research, discussion with
colleagues, and a search of recent BUBL updates bulletins, the
librarian states that he is fairly confident that there isn't.
___________________________________________________________________________
Stage 11 Identified Development Requirements
Client / Server Requirements
(12) Related works search (i.e. more like this title)
Other Requirements
(e) Librarians must have access to information services
which enable them to keep up to date with new public
ations in the electronic world, just as they now have this for the
hard-copy world. These would include publishers'
catalogues, online and printed, and services like BUBL.
___________________________________________________________________________
B.12. STAGE 12: KNOWN URN SEARCH
Having been given a hard copy article which makes reference to a number of
EIOs by quoting, amongst other things, their URNs, the user utilises her
OPAC client to conduct a distributed URN search of the Internet. Depending
on how thing have developed in the interim, the CATRIONA OPAC client would
either search the other Internet OPACs it 'knows about' or would send the
query to the URN to URL resolution service. In either case a number of
alternative URLs would be presented to the user and the CATRIONA client
would act on one or other of these by loading a NIDR client and retrieving
and displaying the EIO.
___________________________________________________________________________
Stage 12 Identified Development Requirements
Client / Server Requirements
(m) Ability to conduct a search for items utilising a URN.
Other Requirements
(f) A URN to URL resolution service perhaps utilising
library catalogues as a stable source of URN to URL
mappings.
___________________________________________________________________________
B.13. STAGE 13: ELECTRONIC REFERENCES
Having retrieved and displayed a multimedia HTML document, the user
decides to access some of the references at the end of the document. These
include a URN for each of the EIO references in question. Depending on how
things have developed in the interim, the user would either: a) 'click
on' the URN and the Web client she is using would access the URN to URL
resolution service and attempt to retrieve and display the EIO utilising
first one, then, if necessary, another of the associated URLs recorded in
it. (or) b) 'copy' the URN back to the CATRIONA client and utilise it to
conduct a distributed URN search of Internet OPACs.
___________________________________________________________________________
Stage 13 Identified Development Requirements
Client / Server Requirements
(n) Ability to copy URN to CATRIONA client and utilise it
to conduct a URN search.
___________________________________________________________________________
B.14. STAGE 14: Z39.50 URL
Having accessed an EIO utilising a WWW client loaded by a
CATRIONA client, the user finds a reference to a Z39.50
OPAC in the document and wants to access and search that
OPAC. The web client 'recognises' the Z39.50 URL and
automatically loads a CATRIONA client to enable searching of
the OPAC to take place.
___________________________________________________________________________
Stage 14 Identified Development Requirements
Client / Server Requirements
(21) Ability to be loaded by e.g. Mosaic (e.g. if Mosaic finds
a Z39.50 URL).
___________________________________________________________________________
Note:
This version of the model does not take account of the possibility
of finding other subject-based OPACs by conducting a WHOIS++ directory
service search as proposed in the ROADS initiative (ROADS, 1995). This
will be taken into account in later versions of the model.
Appendix C.
Methodology and other Background DetailC.1 Information Gathering
The survey and examination of hard-copy and electronic
literature relating to various aspects of the subject area was
carried out throughout the project:
Network navigation issues
Material identifying the problems of location and retrieval of
electronic resources was gathered from various articles and
reports, the most important of which are listed in the
References section.
Investigation of developments in cataloguing standards and
resource description
Much of what is happening in these areas can only be readily
found by means of discussion lists and archived files.
Information on MARC developments was monitored through
the USMARC discussion list and archive. Other lists such as
Autocat were monitored as they occasionally mentioned the
cataloguing of Internet resources. The TEI list, as well as being
a forum for the discussion of TEI development issues, has an
archive of TEI documents. The Internet draft archive and
archive of the IAFA discussion list provides information on
IAFA template development, although the IAFA discussion list
itself has been discontinued. (See Lists in References section)
Uniform Resource Names, Resource Locators, Resource
Characteristics (URNs, URLs, URCs)
The URI list and Internet draft archives were the main source of
material on the rapid developments in these Internet standards.
Most of the discussion on the URI list focuses on specific
technical areas of the various standards, while the Internet
drafts provide more general information. (See Lists in
References section)
Internet cataloguing projects
The aim was to find out about projects that might have a
bearing on the CATRIONA project. Information was gathered
from discussion lists, asking questions on the lists, keyword
searches and searches of hard-copy literature, as well as
enquiries received from those involved in related projects when
information about the CATRIONA Project was posted up on the
BUBL WWW server and on lists. The main related project is
the second OCLC Project: 'Building a Catalog of Internet
Resources' which began in October 1994 (OCLC, 1994),
although there has not been any feedback from the project as
yet. (See Appendix E for a full list of related projects identified).
Z39.50 protocol and implementation
Because this protocol is developing so quickly (Version 3 is
currently under discussion), it was found that most of material
that was available on the subject referred to the previous
versions of the standard and was therefore out of date. Up to
date information was obtained from the Z3950IW list, at
meetings held by the Z39.50 Pre-Implementation Group,
LAITG (Central Scotland) and CIGS (joint meeting), and the
SCOLAR Technical Sub-Committee (See Appendix H) and by
reading the proposed Version 3 standard itself. (ANSI/NISO,
1994). Another useful source was the pointer page of Z39.50
resources available at URL:
http://ds.internic.net/z3950/z3950.html
It has been recognised that there is a lack of easily accessible
background information to the new version of the Z39.50
standard and Robert Waldstein from the Z3950IW discussion
list is currently compiling a new FAQ (list of Frequently Asked
Questions with answers).
Z39.50 Servers and Clients
News about developments in NIDR clients was gathered from
BUBL, various discussion lists and journals.
C.2 Creation of a wide range of hypertext references on the
BUBL WWW and Gopher Servers
A list of hypertext references to some of the electronic resource
materials used in the project was created. Topics covered
include cataloguing, navigation tools and issues, projects,
URN/URL/URC (Uniform Resource Characteristic - see
Glossary) documents and Z39.50.
Available from http://www.bubl.bath.ac.uk/BUBL/catlinks.html
C.3 Development of CATRIONA Model Versions 1, 2 and 3
The CATRIONA model developed from the initial proposal at
URL:
C.4 Discussion of model over e-mail and at face-to-face
meetings with systems suppliers and cataloguing services
Systems suppliers and cataloguing services sent in responses
to the model (Geac, Dynix, FDG, SLS, MDIS, OCLC, RLG)
and these formed the basis for further meetings and email
discussions with systems suppliers and cataloguing services.
Meetings to discuss CATRIONA were held with Geac (October
1994), FDG and CURL (November 1994), Dynix and MDIS
(November 1994). Discussions with other suppliers including
SLS took place at meetings of the Z39.50 Pre-Implementation
Group.
C.5 Dissemination of information relating to the project through
BUBL and e-mail discussion list
Information about the CATRIONA Project was disseminated via
the BUBL WWW and gopher servers. The main documents
relating to the project: the initial proposal, the CATRIONA
model, and the proposal for a second phase, were posted.
(See:
gopher://ukoln.bath.ac.uk:7070/0/Link/Catriona/Catriona_Phase
II)
The model document was also sent to several LIS email
discussion lists including (e.g. PACS-L, LIS-LINK, LIS-BAILER,
Z3950PIG) for feedback.
C.6 Presentations of the model and cataloguing issues
An opportunity was provided to present the CATRIONA model
and to discuss cataloguing issues with cataloguers and systems
staff at a meeting of the Cataloguing and Indexing Group
Scotland (November 1994) which focused on the accessing
and cataloguing of Internet resources and the CATRIONA
model. Aspects of the model relating to Z39.50 were discussed
at the Z39.50 discussion meeting held by the SCOLAR
Technical Sub-Committee and Ameritech (See Appendix H).
C.7 Z39.50 Issues
(See Appendix H)
C.8 Survey, installation and examination of a number of
existing Z39.50 OPAC clients
Several systems suppliers provided Z39.50 OPAC clients for
test evaluation. Geac provided GeoPac Release 1.23, Dynix
provided WinPAC Beta 1.0, and Horizon 3.2, and SIRSI
provided VIZION. Access to the SLS Libertas 6.3 client and the
Innovative Interfaces Innopac Z39.50 client was carried out
over the Internet using Telnet. Demonstrations of the Fretwell-
Downing client and MDIS prototype client were provided during
visits to the companies. WinSpirs (the SilverPlatter ERL client
for access to its ERL servers was also provided for
examination). Data Research Associates also provided their
Z39.50 client but because of time constraints it could not be
included in the evaluation.
C.9 Survey, installation and examination of a number of NIDR
clients
A range of NIDR clients were installed and evaluated including
Cello, MS Windows Mosaic, Netscape, Hgopher and WinWais.
C.10 Testing the Feasibility of the Model
(See Appendix D)
C.11 Design of a Proposed Demonstrator Project
A demonstrator project based on the CATRIONA model was
designed with a view to submitting an 'expression of interest'
under the FIGIT initiative. An updated version of this, amended
to take account of later feasibility study developments, is
outlined in Section 8 of the main report. Alternative designs for
a demonstrator project were not considered, although it is the
case that a range of options exist and that a demonstrator
project could, within certain limitations, be tailored to fit a
desired level of funding.
C.12 Formulation of development requirements for proposed
demonstrator project
A list of development requirements for the proposed
demonstrator model was drawn up and amended during
discussions with suppliers. This consisted of a list of 26 client-
related requirements and 6 server-related requirements (See
Section 6) drawn from an analysis of Version 3 of the model.
The list was then distributed to the systems suppliers during
December with a request for feedback in the first half of
January. The systems suppliers were asked to provide
information on which of the client requirements were already
being developed, which were planned and which were not yet
planned, with approximate time and cost estimates for these
categories. (See Appendix D for results).
C.13 Discussion of Likely Costs of Phase II with Selected
Suppliers
An attempt was made to gauge the likely costs of Phase II.
Because of time limitations, only selected suppliers were
utilised in this exercise. Results are summarised in Appendix D
and in Section 9 of the main report. However, they were
difficult to specify exactly at this stage.
C.14 Preparation of report
The report was prepared with a view not only to presenting the
results of the project but also with the intention of creating a
summary of information and contacts gathered to date.
C.15 Limitations of the Feasibility Study
Some areas which required investigation could not be
addressed in any detail because of time constraints. These
included the following:
Future Z39.50 developments
Links with /conversion to and from IAFA templates and
TEI headers
Access control issues in Z39.50
Improvements to existing tools like Veronica and
Archie
Web client Z39.50 gateways and autoload of
appropriate OPAC client
Appendix D.
This appendix focuses on the feasibility of setting up the
CATRIONA model at a basic level, describes how this was
done, examines the current capabilities of the various OPAC
clients, and briefly describes each client.
Feasibility Tests:
Background Information, Client Capabilities and RequirementsD.1 Feasibility Tests: Background Information
GeoPac provided the main test of the feasibility of setting up
the CATRIONA model at a basic level. It proved to be possible
to set it up so as to be able to access three remote Z39.50
OPACs (Brigham Young University, Butler University and North
Carolina State University).
Brigham Young University
Host=adm4.byu.edu
Port=20003
Database Name=OPAC
AttributeSet=1.2.840.10003.3.1
Host=ruth.butler.edu
Port=210
sDBNames=Marion
Host=library.ncsu.edu
Port=210
Database Name=MARIOND.2 Feasibility Tests: Background Information
The first database searched was the BYU server at Brigham
Young University, in which there were two records with URLs in
the 856$u subfield.
Brigham Young University Library Server (BYU)
BYU Record 1
Logical record length 590
Record Status n
Type of record a
Bibliographic level m
Base address of data 109
Encoding level b
Descriptive cataloging form a
Linked record requirement b
035 bb‡a1234567
100 1b‡aAlberti, Bob
245 14‡aThe Internet Gopher Protocol;‡ba distributed
document search and retrieval protocol
‡c/ Bob Alberti, Farhad Anklesaria, Paul Lindner, et al.
260 bb‡aMinneapolis,‡bUniversity of Minnesota
Microcomputer and Workstation Networks Center
650 b0‡aInternet (Computer network)
856 0b‡agopher.med.umich.edu‡fInternet Gopher Protocol
‡ugopher://gopher.med.umich.edu:70/00/Information%
20About%20Gopher/Internet%20Gopher%20Protocol
Logical record length 298
Record Status n
Type of record g
Bibliographic level m
Base address of data 97
Encoding level b
Descriptive cataloging form ?
Linked record reqirement b
007 kh?buz
008 941102b????????unk???????????bbb???eng?d
035 bb‡a1234568
100 1b‡aJohnson, Charles Ellis.
245 10‡aPhotograph of Joseph F. Smith
856 0b‡alibrary.byu.edu‡fPhotograph‡uhttp://library.byuu/multimedia/jfsmith.gif
Butler University Library Server
Butler University Library has at least 3 records with URLs in the
856$u subfield.
The first record was a full-text HTML format file located at
Indianapolis.
Marc record type USmarc
Logical record length 1131
Record Status c
Type of record m
Bibliographic level m
Base address of data 181
Encoding level I
Descriptive cataloging form a
Linked record reqirement b
001 ABR-9659
008 941007s1994bbbbinubbbbfbbbzbbbbbbbbengbd
099 bb‡ahttp://www.butler.edu/www/library/papers/lotf.html
100 1b‡aKambitsch, Timothy G.
245 10‡aMainstreaming our online catalogs‡h[computer
file] /‡ by Tim Kambitsch
(kambitsch@butler.edu)
260 bb‡aIndianapolis,‡c1994
300 bb‡a1 computer file in HTML format; anchors to 5 GIF
illustrations
520 bb‡aThis paper outlines a model of providing
intellectual access to internet resources through the use of the
World Wide Web. The author advocates that library, system vendors and
database providers make databases available from WWW saavy servers or
through gateways. Librarians can participate in providing
access to specific electronic resources via web-aware online catalogs.
650 bb‡aLibraries‡xAutomation
650 bb‡aInformation retrieval
650 bb‡aLibraries and electronic publishing
856 7b‡uhttp://www.butler.edu/www/library/papers/lotf.html
‡awww.butler.edu‡2World We idWeb
857 7b‡uURL:http://www.butler.edu/~kambitsch/kambitsch.
html‡2World Wide Web
North Carolina State University Library Server (NCSU)
NCSU Record 1
Marc record type USmarc
Logical record length 1540
Record Status c
Type of record a
Bibliographic level s
Base address of data 373
Encoding level b
Descriptive cataloging form a
Linked record reqirement b
001 AFW-1740
003 OCoLC
005 19940616115148.0
008 920721c19929999dcubx1bbbbbbbb0bbba0engbd
010 bbçasnb93004037bço26226155
040 bbçaVPIçcVPIçdNSD
012 bbçla
022 0bça1069-7799
042 bbçansdpçalcd
082 10ça025ç212
090 bbçaZ673.A5çbA42
049 bbçaNRCC
210 0bçaALA Wash. Office newsline
212 1bçaAmerican Library Association Washington Office newsline
222 b0çaALA Washington Office newsline
245 00çaALA Washington Office newslineçh[computer file]:çbALAWON :
an electronic publication of the American Library Association Washington
Office.
246 10çaALAWON
260 bbçaWashington, DC :çbThe Office,çc[1992-
265 bbçaAmerican Library Association Washington Office,
110 Maryland Ave.,.NE, Washington, DC 20002-5675
310 bbçaIrregular
362 0bçaVol. 1, no. 1 (July 9, 1992)-
500 bbçaMode of access: Electronic mail on BITNET;listserv@uicvm;
SUBSCRIBE ALA-WO First Name Last Name
500 bbçaTitle from title screen.
650 b0çaLibrariesçzUnited StatesçxPeriodicals.
650 b0çaInformation servicesçzUnited StatesçxPeriodicals.
610 20çaAmerican Library Association.çbWashingtonOfficeçxPeriodicals.
710 20çaAmerican Library Association.çbWashington
Office
936 bbçaVol. 2, no. 18 (May 10, 1993) LIC
856 bbça
http://dewey.lib.ncsu.edu/stacks/alawon-index.html
D.3 Current Client Capabilities
The functionality of the 6 Z39.50 clients that were made
available or demonstrated was compared. Systems developers
were also asked to indicate any features which were under
development or were planned. The MDIS 'Lion' client was
observed but is not included as it is under development. A
request was also made for the VTLS client and the TRW
SmartSearch client but access to these was not provided. The
focus of the assessment was in relation to the features that
would be required in a CATRIONA type client and so the list of
features follows the sequence of stages in the user's search
described in the model (Appendix B).
Main capabilities required
Notes on features and table entries
As can be seen from the table below, each of the features
required for the model to function at a basic level are available
in at least one of the clients. In some cases an alternative
feature was provided:
If the number of hits is very small, e.g. 1-4, in some
cases (e.g. Libertas 6.3) the full records are
retrieved.
7. Warning if result sets are too large
Rather than providing a warning if the result set is too
large, most of the developers used the alternative
method of setting up the size of result set required in advance.
14. Save search strategy
This only applies to one search and does not mean
"save search strategy from search a and then
combine with search b".
16. Servers to appear in catalogue
The currently available method of offering a choice of
servers is by providing a list for the user. This is
simple to use, but does not give any indication of which server
to select for the kind of search, although VIZION
gave a basic searchable subject listing for the
various servers. One possibility is to create searchable
catalogue records for servers with a subject description.
Another possibility is to use some form of directory service.
20. Close down client session button
This refers to the NIDR client which the user may wish
to close down, and not to the Z39.50 client
which it is assumed can be closed down. Some suppliers
required clarification on this feature.
Table of Z39.50 Client Capabilities
| Features | WinPAC Beta 1.0 | Horizon 3.2 | GeoPac Release 1.23 | VIZION | Oracle Libraries Client | Libertas 6.3 |
| 1. PC client | Yes | Yes | Yes | Yes | Yes | No. Definitely planned. |
| 2. Data for accessing other catalogues not on menu can be uploaded from catalogue | No. Under review. | No. Facility for adding other catalogues under development. | No. But other catalogues can be added if user has data. | No. But other catalogues can be added if user has data. | No | No |
| 3. Distributed (parallel) searching | No | No | No. Under development. | No. Planned. | Yes | Yes. (Serial searching). |
| 4. End-user can select group of servers | No. (Single server). | No. Under development. | No. (Single server). | No. (Single server). Planned. | Yes | Yes |
| 5. Number of hits given before records received | Yes | Yes | Yes | Yes | Yes | Yes |
| 6. Number of hits from servers given separately | No | No | No | No | Yes | Yes |
| 7. Warning if result set is too large | Yes | Yes | No. Size can be set in advance. | No | No | No. Size can be set in advance. |
| 8. Stop button for each server | No. Can't stop search. | No. One stop button. | No. One stop button. | No. One stop button. Planned. | Yes. And overall stop button. | No |
| 9. Option to merge results from several servers | No | No. Possible for user to do locally and manually. | No | No. Planned. | Yes | No |
| 10. Sorting of results (Z39.50 Version 3) | No. Version 3 server function. Client can do with Version 2. | Yes. (Single server). | Yes. (Single server). | No. Planned. Though best done on servers. | No. Planned. | No. Planned. |
| 11. Filing duplicates together | No | No | No | No. Planned. | No. Possible. | No |
| 12. Related works search | Yes. (One server). | Yes. (One server). | Yes. (One server). | No | No. Possible. | No |
| 13. Ability to call up NIDR clients (via URL in catalogue record) | No. Called up via WinGOPHER | Yes. (Not in MARC 856$u). | Yes. (In MARC 856$u). | No. NIDR clients in menu. | No. Near future. | No. Planned. |
| 14. Save search strategy/ results | Yes/Yes. (Single server). | Yes/Yes. (Single server). | Yes/Yes. (Single server). | Yes/Yes. (Single server). | No/No. Near future. | No. Can search same group of servers./No. |
| 15. Levels of access/ charging | No. Planned. | No. Planned. | No. Planned. | No | No. Near future. | Yes |
| 16. Servers to appear in catalogue | No | No. Menu provided. | No. Menu provided. | No. Menu provided. | No. Will be in DALI. | No. Menu provided. |
| 17. Ability to search for a non Z39.50 Internet OPAC | WinPAC has Telnet option. | Can do using SQL. | Yes. Telnet access via a defined URL. | Yes. (Not via URL in catalogue record). | No. Planned. | No. |
| 18. Ability to search for only electronic resources | No | Could if server supports it. | No | No | No | No |
| 19. Ability to handle URLs in other Z39.50 record syntaxes e.g. GRS | No. Planned. | No | No | No | No. Near future. | No |
| 20. Close down NIDR client session button | No | No | Yes. But error message given. | No | No | No. Planned. |
| 21. Ability to be loaded by e.g. Mosaic | No | No | Further investigation required. | No | No | No |
| 22. Ability to offer "other subject catalogues" option after a subject search | No. (Subject list for each OPAC). | No. (Subject and title keyword list). | No | No. (FILTER provides subject search of OPACs) | No | No |
| 23. Offer to search other Internet resources if known item search fails | No | Others can be searched and system would prompt user. | No | No | No | Quick Search option does this. |
| 24. Ability to automatically search other OPACs in turn for known item and stop searching when found | Under review. | No | No | No | No | No |
| 25. Gateway to search ERL-compliant databases | No | Under development. | No | No | No | No |
HORIZON 3.2
A demonstrator version of Horizon 3.2 was provided for testing.
This client provided most of the key features and several
additional features. A number of developments are planned or
already underway, such as remote distributed searching of
more than one Z39.50 site, and the facility to add servers for
searching that are not on the menu originally provided. The
resource can be accessed and downloaded as well as the
catalogue record for the resource, but this is not yet available
according to established standards (i.e. the URL is not in
MARC 856$u).
GeoPac Release 1.23
GeoPac provided most of the key features required as well as a
user-friendly interface and was successfully used to access and
download electronic resources in remote locations using a
standard method of resource description. Additional servers not
on the menu can be added and searched if the user knows the
connection details. Distributed searching is currently under
review.
VIZION
A test version of VIZION was provided for evaluation. The
interface is fairly user-friendly although the user has to work out
which choice to make first rather than being guided. VIZION
offers a searchable list of destinations including Z39.50 sites,
Gopher sites and WAIS sites. The user chooses and the client
uses the appropriate protocol to communicate with the server.
Servers can be added if the user knows the connection details.
ORACLE LIBRARIES CLIENT
The Oracle Libraries client is capable of distributed searching
and this was demonstrated. The Oracle Libraries client has
already been up and running for some time (See Appendix E -
Projects for information on the IRIS project). The facility for
distributed searching was the main advantage of the Oracle
Libraries client, although the distributed searching was not
carried out by the client itself but by a Unix 'collator' which sent
off the requests to the 5 Irish servers, merged the results and
sent them back to the client. The resources themselves could
not be accessed or downloaded but only the records.
Libertas 6.3
This client differed considerably from the other clients in that it
is not a PC client, although SLS plans to bring out a PC client
in 1995. The other main difference was that this client is aimed
more at library staff rather than the end user and offers a wide
variety of facilities for downloading records in various formats
and modifying records. It is understood that the PC client will
be aimed more at the end-user. Again the main advantage of
the SLS client was the facility to do distributed searching. The
servers were searched serially which made the response time
longer, especially for the US servers. Again the resources
themselves could not be downloaded - only the records.
___________________________________________________________________________
ALIWEB - Archie-Like Indexing in the Web
CNI TopNode
Harvest
INFOSERVICES
NISS Information Gateway
OCLC: Building a Catalog of Internet-Accessible Materials
OMNI - Organizing Medical Networked Information
SOSIG - Social Science Information Gateway
LYCOS
Subject Trees
WebAnts
WWWW (World-Wide Web Worm)
DALI (Document and Library Integration)
DECOMATE (Delivery of Copyright Material to
End-Users)
DVB-OSI II
EUROPAGATE
ION
IRIS (Irish Networked Database Service)
NORDIC - SR-Net
SOCKER
Sources: Lorcan Dempsey (1994) and see also Appendix H;
Fretwell Downing; (VINE, 1994) and information available at
the above URLs.
___________________________________________________________________________
This is a problem that has not been addressed by those
involved in resource description until relatively recently.
However, a number of initiatives are now underway to
accommodate the cataloguing of EIOs, conducted both by
those traditionally concerned with cataloguing standards,
MARBI in particular, and by those newer to the field such as the
URI (Uniform Resource Identifier - See Glossary) and TEI
groups.
In this section we consider some of these initiatives from the
perspective of what is required for the CATRIONA model to
operate successfully, looking both at what has been achieved
and what remains to be done.
It should be stressed at the outset that almost all of the
enhancements to resource description required for the
CATRIONA model to function at a basic level are already in
place in one of the cataloguing schemes (USMARC). The other
three schemes seem to be moving in this direction with IAFA
templates coming closer to fulfilling the CATRIONA
requirements than URCs or TEI headers, although all three
schemes are presently inadequate for CATRIONA since, at
present, they are not able to be handled by Z39.50 clients and
servers.
The basic level requirements of a cataloguing scheme needed
for the CATRIONA model to function are listed below as (a) (-
to a certain level), (d) and (f). The remaining requirements are
not necessary for the demonstrator model to function, but
would be necessary for a fully functional, integrated
implementation of the CATRIONA proposals.
BASIC REQUIREMENTS
(a) Ability to catalogue any and every possible type of Internet
resource and service
ADDITIONAL REQUIREMENTS FOR FURTHER MODEL
DEVELOPMENT
(b) Ability to record username and password of commercial
services in suppressible (MARC) field.
In each case the information on the requirements (a) to (j) was
sought in the key documents for each resource description
scheme. In many cases it was not possible within the timescale
of the project to determine whether or not the requirement was
met. This does not mean that the requirement cannot be met,
but only that it has not as yet been possible to obtain clear
information on the question.
Together with other additions, the above enhancement to the
USMARC record format means that it meets the CATRIONA
requirements sufficiently for the model to function at a basic
level. There are several advantages to using the MARC format
for describing EIOs at present and for the near future: it is
already widely used and so cataloguers are not required to
learn a new system; it provides a good standard descriptive
format, provides subject indexing, and it also works well with
Z39.50.
There is the problem of differing national MARC formats to be
considered as these enhancements were developed initially for
USMARC and implementation in other MARC formats is very
limited as yet.
BASIC REQUIREMENTS
(Note: the following description applies to the USMARC format
up to Proposal 95-1)
Resources identified by MARBI as requiring to be
catalogued:
OPACs, Bulletin Boards, Mailing List Servers, Data Archives,
Computational Resources, White Pages, Network Information
Centres, Full-text Databases, Numeric Databases (MARBI,
1991a), other types of citation database (MARBI, 1991b),
Computer Discussion Lists and Forums, FTP Sites (since these
are rapidly changing and numerous, it was questioned whether
they should be controlled bibliographically), General Online
Services (e.g. Prodigy, Compuserve), Campus Wide
Information Systems, Distributed File Servers (Gopher, WAIS,
World Wide Web) (MARBI, 1993b)
ADDITIONAL REQUIREMENTS FOR FURTHER MODEL
DEVELOPMENT
Various efforts are also underway to map between MARC and
TEI headers (See below) which were partly modelled on MARC
(Gaynor, 1994).
Although IAFA templates are a recent development, they are
already being used, particularly in the ALIWEB (Archie Like
Indexing in the Web) retrieval service (See Appendix E), and
as the basis for PICA resource descriptions (See Appendix E
and Dempsey, 1994). However, IAFA templates are not at
present a Z39.50 registered syntax and cannot be used with
existing clients and servers.
IAFA templates are based on attribute pairs, with the number
of attribute pairs and the content of some of the values is left
up to the encoder, which may lead to retrieval problems as the
encoder may well be inconsistent e.g. in choice of keywords,
although the intention is to restrict the syntax and semantics
used to that the templates can be automatically collected and
indexed (Weider & Deutsch, 1994).
BASIC REQUIREMENTS
IAFA Template types
(d) Records must be in a form that can be searched and
retrieved using Z39.50.
ADDITIONAL REQUIREMENTS FOR FURTHER MODEL
DEVELOPMENT
In this section the URC is considered as a means of encoding
meta-information about a given network resource (ibid). There
is also a proposed URC service, in which URNs are resolved
into URLs (Daniel, 1994b).
The URC as it is at the moment is designed to be simple to
understand and encode, extendible, compatible with a wide
range of systems and with existing technology. (Mealling,
1994). The disadvantage of this approach is that its simplicity
and flexibility - records can be encoded by anyone, with the
content of the title and abstract values left up to the encoder -
may present retrieval problems.
Like IAFA templates the URC is based on attribute-value pairs.
See Note 3 for the current minimal set, "although no attribute-
value pairs are required to be part of a URC" (ibid).
BASIC REQUIREMENTS
ADDITIONAL REQUIREMENTS FOR FURTHER MODEL
DEVELOPMENT
TEI headers have advantages in that they are flexible - there
are very few mandatory elements, and they make use of the
standard language SGML. There are potential
retrieval problems, however, as the values are mostly open
(ibid). Other limitations for use in the CATRIONA model is that
access via URL is not yet accommodated, and the resource
descriptions required in the CATRIONA model must cover a
wide range of systems and services as well as electronic texts.
BASIC REQUIREMENTS
ADDITIONAL REQUIREMENTS FOR FURTHER MODEL
DEVELOPMENT
A GILS Core will be set up including records for all information
locators that catalogue publicly accessible government-funded
information resources. It is expected that the Core will be no
more that 100,000 locator records. The locator records will
conform to the GILS Core Element standards. The definitions
of the Core Elements identify the corresponding USMARC field
for each element. Listed below are the mandatory elements as
well as optional elements of relevance to the CATRIONA
requirements.
___________________________________________________________________________
NUL uses the Dynix library automation system. During the
selection process for this system, the Depute Librarian wrote to
the Technical Director of Dynix Library Systems (UK) Ltd at the
beginning of 1990 regarding development of the system:
The Library wishes to develop the existing environment ... to
attain the following:
To some extent, the first development has now been attained,
and the latest Dynix upgrade (currently planned for installation
at Napier towards the end of 1995) allows graphical images to
be attached to standard MARC records. Editing and uploading
facilities are not currently well-developed. The second
objective was reached by using a shareware Windows terminal
emulation package, which allows multiple Dynix sessions to be
displayed on the same Windows desktop, and for data to be
moved between them using the Windows Clipboard. The third
objective is under development by Ameritech Library Services,
the current owners of Dynix, using a Windows environment.
The Strategic Plan 1994-98 of NUL contains proposals which
include:
Library emphasising service to clients, but trying to increase
their self-help and self-service capabilities
Considered mixes of printed, audiovisual and electronic
information provision and of policies for holdings and access
Electronic access ... to research information from JANET and
the Internet, University-generated teaching packages and
University administrative, advisory and regulatory information
This screen integrates access to several resources:
Title 1 is a book containing extracts from the newspaper.
Selecting any particular title results in the display of more
detailed descriptive information in card format, as in Screen 2.
Location and availability status is displayed on a subsequent
holdings screen. The CD-ROM versions have no physical
format because they are mounted on the LAN, and therefore
have no particular campus location and are always available.
There are no data links between the MARC records in the
catalogue and the electronic files because Dynix cannot
support them; instead, the link is made explicit to the user in a
dummy holdings record, as in Screen 3. This information could
also be incorporated as a local note, MARC tag 509, in the
bibliographic record; it has to be attached to the holdings record
in this instance because it does not apply to the other holdings
which refer to the manuals.
Users who wish to access the CD-ROM must log-off from the
catalogue, and log-on to the University Network CD-ROM
Menu, from where the title can be selected. When finished with
the CD-ROM, users must log-off from the CD-ROM Menu, and
log-on to the catalogue if they wish to continue searching for
Library materials.
In addition to networked CD-ROMs, NUL is having to cope with
an increasing number of other electronic resources, held locally
on the LAN or available via JANET. These include DOS
abstract and indexing services, multi-media packages, BIDS,
and the OPACs of other libraries. In all instances, dummy
records have to be used in the catalogue, and users have to
carry out a cumbersome switch between different systems to
access the resources they have identified.
For example, it is expected that the Guardian will introduce an
online version of the newspaper to be accessed over the
Internet/JANET, in the very near future, as several other
newspapers have done. It may also be the case that another
organisation offers a service collating literary and arts reviews
from several UK national newspapers, including the Guardian,,
and mounts the results on the Internet. Other hypothetical
products could include a CD-i on editing a newspaper, using
the Guardian as the example, and an online file of images of
press photographs, etc. NUL would attempt to integrate
identification of these extra resources by creating catalogue
records similar to the CD-ROM resource, with associated
dummy holdings and explicit connection details, as in Screens
4 and 5.
In this case, location and availability information is incorporated
into the MARC record; the dummy holdings record would be
similar to the existing example. The user is still required to log-
off from the catalogue, having noted the http address, and log-
on to a University JANET/Internet browser service. If the user
subsequently needs to check the catalogue, say for an original
printed copy of an issue of the newspaper to see its typography
and page layout, then the whole process has to be explicitly
reversed.
Another anticipated problem with this method is the
proliferation of standard instructions and location codes needed
to given explicit access information. Access may be via one of
several different direct, dial-up facilities or pre-set menu
systems supporting World-Wide Web, Gopher, Telnet, etc.
A further problem with providing explicit connection information
is that it may change. The catalogue record will require
maintenance to ensure the correct URL or other network
address, whatever model evolves. The user may, however, try
to use an old address, noted during a previous catalogue
search, without checking the latest information. This is
analogous to a user going direct to the shelf for a book
previously consulted, without checking the catalogue, only to
find the item has been reclassified, respined, and transferred to
another branch library.
The preferred approach for NUL is thus a system which allows
users to identify all types of resources held locally, using a
common search interface, and to extend the search to
resources held by other libraries without reformulation. If the
identified resource is electronic, the user should be able to
access it within whatever license arrangements have been
made, by proceeding directly from the catalogue record without
having to reverse and take another route. Once use of the
resource is complete, the user should find themselves back at
the catalogue record, ready to proceed to check other retrieved
resources, or to carry out another search.
In other words, disregarding any other aspects of a graphical
user interface, NUL would like to offer its users Screen 6 rather
than Screen 5.
Screen 6: The active catalogue record
DYNIX BIB # 999999
TITLE The Guardian online
The Guardian
The Observer
NAME(S) 1) Guardian Newspapers Ltd
PUBLISHER [London] ; Guardian Online Ltd ;
SUBJECT 1) 072.1
ACCESS Click on button to use this service
- - - - End of Title Info - -
- -
___________________________________________________________________________
The principal speaker, on Z39.50, was Lorcan Dempsey,
Director of the UK Office for Library Networking.
Lorcan began his talk by noting the growth of the network
environment, the existence of location devices such as URL
and URN, and the importance of meta-data such as finding lists
and library catalogues. Other features included client/server
system architectures, and a number of different resource
spaces such as Telnet, WWW, Gopher, and FTP. Although
these spaces are becoming linked, most network resources of
interest to libraries used Telnet, which has many shortcomings.
Increasing connectedness between systems or resources was
resulting in a lot more data flowing around networks.
The library resource space is currently characterised by
monolithic applications using terminals as the main peripheral
device. It is fragmented, and navigation tools are small-scale
and not as elegant as they might be (clunky).
The establishment of the Z39.50 standard is an attempt to
resolve some of these problems. It is a protocol for translating
a local database search request into a standard form which can
be interpreted by a foreign database into its own searching
procedures, so that the original request does not require
rekeying or reformulation. Although it currently focuses on
searching and retrieving bibliographic information, it is intended
to be generalisable to non-bibliographic datasets such as
statistics, chemical formulae, etc. The results of a Z39.50
search are always in the form of a set of records which may be
in the format, for example, of a standard OPAC card image.
A number of standard Z39.50 attribute sets are evolving, to
cover different datasets; these include bib_1 for bibliographic
data, GILS for US Government information, and STAS for
chemical data. These attribute sets define search types, such
as Title, and operators, such as =, which can act on them.
Several different record syntaxes are currently catered for,
including MARC, SUTRS (for simple text), and the
aforementioned OPAC (simple bibliographic data with some
holdings and circulation status data).
Projects proposed, underway, or completed, using Z39.50
include DBV-OSI2 in Germany, IRIS in Ireland, LONDON
LINK, and ALSA. The last had been proposed for FIGIT
funding by UKOLN, but had been turned-down. The British
Library is working on a project for a general OPAC to European
national libraries, and a lot of work is being carried out in the
Netherlands by PICA.
The following points were made in summing up the current
situation:
Z39.50 itself is only a part of the network picture.
Distributed library systems are in an immature state.
There is no driving focus in the UK for a distributed, virtual
library.
There is no culture of resource-sharing in the UK, an essential
component of the network approach.
There are no significant UK resources available to improve this
situation.
In response to questions regarding the enhancement of Z39.50
searching to provide search control structures such as menus
for databases to be searched, time-out facilities, and automatic
search halting if a hit was made on a unique key such as ISBN,
it was said that little progress was being made, although
implementors were aware of such issues.
After a coffee-break, the meeting was given a demonstration of
the SALSER system by Bruce Royan, Director of Information
Services at the University of Stirling. SALSER is a finding tool
for serials holdings in Scottish libraries. It essentially consists
of a Union list of serial titles, with brief location and holdings
information. It has been designed so that the SALSER serial
title can be linked to the full catalogue of the holding libraries
using Z39.50, so that issue detail and availability can be
displayed to the user. Eventually, the whole system may be
supplanted by a general Z39.50 interface, but this will not
happen until every member library has an automated serials
catalogue, with an appropriate Z39.50 client.
Status reports of Z39.50 implementation were given from the
British Library, Manchester Computer Centre, Specialist
Computer Group, Bangor University Library, Fretwell Downing
and Strathclyde University Library. The current status of the
CATRIONA project was reported on and there was an
opportunity to present the aims of the project in general terms.
Fretwell Downing gave a description of the related IRIS and
DALI Projects (See Related Projects and Developments). The
work of EWOS EGLIB was then presented by Erik Lorenz
Petersen, and was followed by a discussion in which concern
was expressed at the need for progress of European
information retrieval standards development in relation to what
is happening in the US.
It was decided to host a Z39.50 seminar in January 1995
(Z39.50 UK Wide: What's Going On?) in order to inform the
LIS community of what is happening in Z39.50 implementation
in the UK.
The major part of the meeting was allocated to systems
suppliers who demonstrated various clients or outlined
development plans. SLS gave a demonstration of Libertas 6.3,
GEAC gave a demonstration of GeoPac, and Bryan Alton of
IRIS Document Delivery Services Limited gave a
demonstration of the IRIS client. This was followed by several
updates on development plans. The MIRO System Architecture
for searching online hosts and which uses Z39.50 was outlined
by Satellites International. Neil Smith gave a presentation of
the British Library plans to develop a Z39.50 server and
discussed the ION project. The DALI project and the future
trend towards providing services to the student desktop was
outlined by Fretwell-Downing, as well as the need for LIS
professionals to become involved as gatekeepers. Peter Smith
of LASER discussed various aspects of their projects, including
London Link. The position of BLCMP was described by Robert
Watson who said that they may make their BLCMP database
accessible via Z39.50 if there is sufficient demand.
The meeting closed with an observation from Lorcan Dempsey
that the proposals for the Libraries Programme in the EC
provide evidence that there is a need to bridge the awareness
gap between those who know what the technology can do and
those who are responsible for providing funding. He indicated
that UKOLN would be interested in supporting such measures,
which might also involve the Z39.50 PIG.
John Kolman then discussed Z39.50 in more depth, detailing
the various requests and responses involved in a Z39.50
session. The potential interoperability problems between clients
and servers which can arise due to differing combinations of
search types, attribute sets and record syntaxes selected were
described. Finally the developments proposed for Z39.50
Version 3 were mentioned including the record syntaxes
SUTRS and GRS, as well as the Explain facility which allows
an origin to obtain details of the implementation of a target,
including databases available for searching, attribute sets and
record syntax.
___________________________________________________________________________
However, it is seen at its best if accessed by a World-Wide
Web client. The URLs for the service are:
One aim of the initiative was to provide the academic
community in the UK with improved access to Internet services
and resources. It demonstrably improves the user's interface
with these resources and so improves the quality and efficiency
of the access. The project has other aims which are equally
innovative:
The project seeks to meet these aims by involving the LIS
community in the creation and maintenance of the Subject
Tree. Although the main tree was largely created by the BUBL
team, they were also assisted by a large group of subject
specialists distributed across the UK (and beyond). This group
also take responsibility for the maintenance and further
development of some parts of the tree and other specialists are
being sought to cover those subject areas not already covered.
They communicate with each other and with BUBL (and NISS -
see below) via the Mailbase electronic mail discussion list LIS-
Subjects. The Subject Tree Initiative is, in fact, only one of the
many ways in which BUBL seeks to meet its aims. The primary
aim of the service is top encourage, develop, co-ordinate and
support the emerging LIS networking community in the UK and
to promote its interests.
It is a fundamental assumption of the BUBL enterprise that its
various sub-projects are examples of good practice. The
Subject Tree Initiative is a good illustration of this, the following
characteristics of it being regarded as examples of good
practice, embodying principles which are capable of being
applied in other contexts:
The following description of the Subject Tree Initiative should
give a clearer picture of the initiative and how it works:
The Subject Tree is based on the Gopher and World-Wide
Web software technologies. These enable BUBL to provide
access both to files held on the BUBL machine itself and, more
importantly, to other remote services and resources. Link
information like the following:
GOPHER
WORLD-WIDE WEB
allows the user to jump from an item found under a particular
subject heading to either a file held on BUBL or a service or
resource anywhere on the Internet. The task of the subject
specialist is to discover valuable resources and transmit either
the resource itself or the link information to BUBL (and also
NISS) with an indication of which part of the Subject Tree it
should go under. The use of gopher and WWW links allows
items to be 'pointed at' from a number of subject areas thus
solving the difficulties of resources of interest in a number of
subject disciplines.
Resource discovery takes place via a number of mechanisms:
Once a resource is discovered, either the resource itself or the
link information is sent to the Mailbase list LIS-Subjects from
which it is picked up by BUBL and also by NISS.
http://www.bubl.bath.ac.uk/BUBL/Tree.html
This information will be added to and enhanced by the BUBL
Information Service as information vital to the library
community and its users regardless of whether or not the
CATRIONA project continues into Phase II and beyond.
___________________________________________________________________________
GEAC Computers Limited
___________________________________________________________________________
The Workshop was structured in two parts: a short presentation
by the Workshop leader; followed by an open discussion on the
issues raised.
Many libraries purchase electronic resources such as CD-
ROMs for local use; these may substitute for printed material.
Such resources can be catalogued in an integrated fashion with
other resources; the information contained is the primary
aspect, not the physical format. Nowadays, however, we also
have to contend with networked information - virtual resources
which have no physical manifestation. Existing cataloguing
rules blur the distinction between intellectual works and their
physical format because the mode of use or access may very
well affect the intellectual structure of the work. Current rules
are weak when physical manifestation is nebulous.
A second problem is that Internet and other networked
resources have no fixed location. They can exist in more than
one location, or the location can change over time.
Maintenance of the catalogue record becomes complicated.
The principles of ownership, copyright and access are also
more difficult to apply under these circumstances.
Electronic resources can be direct equivalents of their analogue
forms, including text, graphics, and sound. To access machine-
readable text, hypertext and multimedia software, users require
programs such as viewers and browsers, often specific to the
resource. Such programs may give users the means to alter a
document - electronic resources are particularly amenable to
change.
Documents need no longer be flat; as well as being non-linear
or multidimensional, they may no longer represent a single
intellectual concept.
Electronic resources can be primary, secondary or tertiary, in
the traditional sense. An electronic resource may be a work, an
abstract of the work, or a catalogue of records of many works.
Electronic catalogues of catalogues are urgently required on
wide area networks; the electronic manifestation of Russells
catalogue of catalogues which reference themselves is just
around the corner.
The existing international cataloguing standards of AACR,
MARC, and ISBD have evolved in a print-based culture. While
they are adequate for describing intellectual works, they have
shortcomings in an electronic environment. ISBD has little to
say on displaying bibliographic information in graphical user
interfaces, AACR lumps all machine-readable files together,
and MARC requires modification to store data associated with
electronic information objects (EIOs). Additional MARC tags
and subfields are needed to store the electronic address of an
EIO, analogous to the shelfmark of a physical resource.
Information about access, including passwords, log-ons, and
editing and copying rights is much more resource-specific than
for traditional materials, and should also be stored as part of
the bibliographic record. The implicit access route for a
physical library may be to join the library and obtain a readers
card; the electronic equivalent is likely to be different
depending on the particular EIO, and must be made explicit to
the user to avoid the need for professional mediation.
A trend in automated catalogues is increased linkages between
records. Implicit links are a primary feature of bibliographic
catalogues, in the shape of authors, titles, and subjects.
Explicit links are now seen as desirable for improving access to
and integration of information; these links can be between
analytic entries and the principal bibliographic record, or
between electronic abstracting and indexing products and
bibliographic and holdings records. Analytics share the lack of
a separate physical presence with EIOs; they are embedded in
other works. Hyperdocuments may consist of intellectually
separate works linked together electronically; again, concepts
of intellectual ownership have to be revised. Such links may
be dynamic, and may be amended automatically by intelligent
software; even the concept of a work needs reappraisal.
The electronic catalogue of electronic information resources
thus faces problems concerning description, location, access
(to the resource), content, ownership, version (or edition), and
integration, all of which are subject to a much greater degree of
change than for traditional resources.
The main advantage of such a catalogue is its own
manifestation: it does not need to be physically integrated to be
electronically integrated. The records do not have to be stored
in a single data file; they can be brought together from many
different files to create a virtual catalogue. From the users
point of view, it does not matter whether the resource is
accessed in China or Canada, and it is of no interest that the
catalogue record being used for access is itself stored in
Dundee or Dresden.
The CATRIONA Project was mentioned; this aims to produce a
catalogue that integrates resources of all types and provides a
navigation tool to identify EIOs. The catalogue approach
reduces netsurfing and prevents time-wasting; on the one hand
there is no need for random browsing (surfing), and on the
other access is controlled by the catalogue record itself.
It was pointed out that there is confusion between the catalogue
and deskspace. The same PC or other microcomputer could
be used to access the OPAC and to download and manipulate
the actual EIO. In some circumstances, users would not be
able to carry out a quick search on the OPAC for local, physical
materials because all machines were being used to prepare
bibliographies. There might therefore be a need to consider
disabling the retrieve facility on some OPACs so that they were
dedicated to standard catalogue searches only.
An example of the desire for a fully integrated catalogue was
The Guardian newspaper: many libraries would already have
records for the newspaper itself, collections of essays from the
paper (The bedside Guardian), books about the newspaper,
illustrations of typography and page layout taken from the
newspaper, and the publishers annual report and accounts. To
this could soon be added records for the newspaper on-line, an
abstracting/indexing service for the paper, a directory of
journalists and other staff employed on the paper, etc. Any
library user seeking information about the Guardian would
expect to retrieve catalogue records for all of these resources
in a single search.
There was discussion on what changes might be required to the
MARC standard; the USMARC initiative on the 856 tag was
mentioned, together with the OCLC research report on the use
of MARC for cataloguing networked resources.
There was a general discussion on the role of the librarian in
the electronic information environment. Librarians could
provide information by storing and providing access at a local
level (catalogue only what you own), by providing specific,
detailed access only at a remote level (catalogue what you
think the user will be interested in), by providing general, non-
specific access (just the access channel, with little navigation),
or a mixture of all modes.
There was potential confusion between communication and
information; opening a communication channel was not the
same thing as using it to transmit information. The workshop
agreed that the librarians skills encompassed both; information
requires expert organisation, but communication factors had to
be taken into account in designing the user interface.
Alternatives to CATRIONA, or any system using traditional
cataloguing methods, were discussed. It was pointed out that
traditional computer solutions, such as menus and icon groups,
worked well when only a few tens of alternatives were to be
organised, but these methods failed when thousands or millions
of separate items had to be categorised. Brute force methods
of keyword indexing would not be sufficient for a user to
identify a particular resource because of the vagaries of
language and the perversity of information creators. Controlled
language approaches, such as authority files and subject
indexing, were necessary to improve precision.
It was agreed that a major problem in cataloguing remote EIOs
would be a change of home location; catalogue records would
then point to the wrong address and the resource would not be
retrievable. Cataloguers would not have the time to constantly
check the currency of Universal Resource Locators (URLs) or
other addresses. An automated process could identify invalid
URLs, but would not be able to determine the correct URL. It
was suggested that one solution might be for EIOs to be
catalogued in the same locality as they were created; if the
creator changed the URL, they would notify a colleague
cataloguer to update to record. Other libraries would see this
catalogue record in their own OPACs, if they chose to do so,
although it would actually be stored remotely, and brought to
the local OPAC by a Z39.50 or similar search. The alternative
approach of copying the record to the local catalogue would not
resolve the original problem of a URL changed elsewhere.
Such methods would require a far greater level of organisation
of the Internet, including rules for network publishing, resource
cataloguing, and location co-ordination than was currently the
case.
Librarians would not be able to achieve this on their own;
practical co-operation between libraries is much talked-about
and little acted-upon. The workshop thought that the best hope
was for one or more central agencies to organise such work:
advising creators on CIP for EIOs, identifying and cataloguing
important resources of general interest, and making such
catalogue records available to libraries. OCLC was already
moving in this direction. Librarians would still have to do it for
themselves, a la CATRIONA, for specific resources of narrow
interest.
The workshop also identified a need for better training in
information exploitation by providers, mediators and users.
And a final observation: only one library represented at the
workshop had actually attempted to catalogue networked
resources, and that was confined to resources such as CD-
ROMs mounted on the local network. While the proportion of
such resources to total stock remained low, reasonable
navigation and access could be achieved using non-integrated
menus or printed documentation, but as their numbers
increased, libraries would have to tackle these problems head-
on.
Appendix E.
Related Projects and DevelopmentsE.1 Template/Record based Projects
Alex: A Catalogue of Electronic Texts on the Internet
The Alex catalogue is a central Internet OPAC for the location
and retrieval of the full text of documents on the Internet. It
currently indexes over 700 books and shorter texts by author
and title, including texts from Project Gutenberg etc. At the
moment it does not have serials.
Access details:
gopher://rsl.ox.ac.uk:70/11/lib-corn/hunter
Alex has recently been converted to MARC format which is
available for downloading.
Access details: ftp://ftp.lib.ncsu.edu/pub/stacks/alex/alex-
950224-marc.txt
The ALIWEB framework is a method of providing distributed
indexing in the Web and thus reducing the load on network
resources. Server administrators complete index files in the
form of IAFA templates (See Appendix F), which are regularly
retrieved and compiled into a database. The database can be
searched using a simple WWW search interface or via other
databases.
Access details:
http://web.nexor.co.uk/public/aliweb/aliweb.html
CNI TopNode is an Telnet-accessible database of directory
level and individual resources in the Internet and is produced
by Merit Networking and Indiana University. The topnode
database has been developed using BRS/Search software on
CNI's server. Individuals and organisations offering a resource
are required to complete registration templates developed for
different types of resources.
Access details:
gopher://gopher.cni.org:70/11/cniftp/projects/topnode
Harvest is a system for efficiently gathering and distributing
indexing information. It supports the construction of various
types of indexes which can be customised to suit the each
information collection; and it provides caching and replication
to prevent network congestion. Harvest is made up of several
subsystems including the Information Gatherer and Information
Broker - which can be searched using author, keyword, title or
URL - as well as subsystems for indexing/searching, replicating
and caching. The indexing format used by Harvest Summary
Object Interchange Format (SOIF) is similar in some respects
to IAFA templates (Bowman et al, 1994). Access details:
http://rd.cs.colorado.edu/harvest/
INFOSERVICES is a collaborative project between SURFnet,
the Dutch Academic Network and the National Library of the
Netherlands to improve access to networked information
resources. Suppliers of information are required to provide
elementary descriptions in an accompanying INDEX file which
is based on the IAFA template standard (See ALIWEB and
Appendix F). The IAFA template material is then enhanced to
form catalogue records for electronic resources. Cataloguing
enhancements recommended by MARBI, such as the addition
of USMARC field 856 were adopted by the project and
implemented in the Dutch Pica cataloguing environment which
allows materials to be accessed through the Dutch OPACs as
well as through the combined central Union catalogue and ILL
system (van der Werf, 1994). In association with OCLC, PICA
have recently set up an experimental Z39.50 interface link to
OCLC's FirstSearch service. This is currently undergoing
evaluation in six Dutch libraries.
The NISS Information Gateway brings together all of the
existing NISS services and employs an enhanced, easy to use
WWW interface. Resources are classified according to a
subject tree approach and described by means of templates.
The templates contain object information, such as information
name, information type, UDC number and free text description;
and implementation information, such as URL, service type,
network address and currency.
Access details: http://www.niss.ac.uk/
In the new OCLC project - Building a Catalog of Internet-
Accessible Materials, libraries and representatives from their
host institutions will identify, select and catalogue EIOs
accessible through the Internet. The records produced in the
project will be made accessible over the Internet. The main aim
of the project is to test and evaluate the effectiveness of using
the USMARC format for bibliographic records, and field 856 in
particular, as a means of providing description, location and
access information for EIOS. The project duration is October 1,
1994 to March 31, 1996.
Access details:
http://www.oclc.org/oclc/man/catproj/catcall.htm
OMNI is a project to build a gateway for access to quality
health and biomedical information resources to the UK higher
education and research community. The gateway will be a
WWW interface to a catalogue of biomedical and health-
related information. Resources will be filtered, classified,
subject indexed and catalogued using a modification of the
NISS/SOSIG template, which underlies the search mechanism.
Access details: http://www.nimr.mrc.ac.uk/OMNI/
SOSIG is an ESRC-funded project to provide a centralised
means of access to social science resources available over the
Internet. All of the resources on the SOSIG server are
catalogued using a standard template (See NISS Information
Gateway).
Access details: http://sosig.esrc.bris.ac.uk
E.2 Other IR Projects
ICE
The ICE indexing server extension enables sophisticated free
text searches on a Web archive. The indexer provides
relevance feedback, use of a thesaurus to make conceptual
searches for e.g. image, forms searching and boolean
searching. As well as being used on individual archives the ICE
software can be used as a data gathering script to provide and
index file for archie like indexers such as ALIWEB. (Neuss and
Hofling, 1994)
Access details: http://www.dsi.unimi.it/cgi-bin
LYCOS currently provides one of the largest indexes to the
Web. The huge index is built by a Web crawler using a system
that combines random choices with certain preferences.
Information gathered includes title, headings and subheadings,
weighted list of 100 terms, first 20 lines, size and number of
works. The index can be searched using document title,
headings, links and keywords and an extract can be examined
which allows the user to evaluate the document before retrieval
(December, 1994).
Access details: http://lycos.cs.cmu.edu
Subject Trees and subject directories are used
for organising subject-related resources in a hierarchical manner.
Examples of these include the following: the BUBL Subject Tree (See
Appendix I. Access
details:http://www.bubl.bath.ac.uk/BUBL/Tree.html or
http://www.bubl.bath.ac.uk/BUBL/Treealphabet.html; EINet
Galaxy (Access details: http://www/einet/net); the WWW
virtual library (Access details:
http://info.cern.ch/hypertext/DataSources/bySubject/Overview.html);
GENBB (Access details:
http://www.cs.colorado.edu/homes/mcbryan/public_html/bb/su
mmary.html); YAHOO (Access details:
http://akebono.stanford.edu/yahoo/). Users can
contribute to these systems by supplying the appropriate information about
their resources.
Foreseeing that single indexes of the Web can only grow at the
expense of coverage, the aim of the WebAnts project is to
develop a cooperative Web explorer which shares results with
other explorers and avoids duplication of effort. The distributed
approach which involves smaller cooperating search engines
means that sites do not have to have large resources available,
and so more sites can be involved. It is also envisaged that a
cooperative method can be used to provide index information
to the end-user.
Access details: http://thule.mt.cs.cmu.edu:8001/jrrl-
space/webants.html
The aim of WWWW, released in March 1994, is to locate all of
the WWW addressable resources on the Internet and provide a
powerful user interface to the resources. The WWWW program
scours the Internet locating all web resources. The search
engine is accessed on the WWWW home page and uses forms
and searches, including keyword searches, can be carried out
on titles, reference hypertext or within the URL name strings of
documents (McBryan, 1994).
Access details:
http://www.cs.colorado.edu/home/mcbryan/WWWW.htmlE.3. European Z39.50/SR Library-based Projects
British Library/ONE
The British Library are developing a target Z39.50 server based
on the NLC code which is due to be available in August 1995. A
range applications of Z39.50 are also being investigated under
the Initiatives for Access projects. The British Library is one of
the 9 partners in ONE, an EU project which includes partners
from Finland, Norway, Sweden and Germany. One of its key
aims are to produce a public domain origin and target kernel
code running over Internet and OSI telecommunication protocol
stacks.
The DALI (Document and Library Integration) project led by
Fretwell-Downing will build on the IRIS project to develop, pilot
and test a multimedia document delivery service in a
distributed environment over Z39.50. One of the key aims is to
develop a modular software package which will incorporate SR
and Z39.50, X.400 and MIME. Other features include a current
awareness service, directory service (X.500) and charging
mechanisms.
The aim of this project is to develop software which will provide
links between bibliographic records and full text articles,
allowing the user to view the article or have it delivered.
DVB-OSI II aims to achieve transparent unified access to a
range of resources most of which are Union catalogues,
including the Pica Union catalogue, but also including
abstracting and indexing services.
EUROPAGATE is a gateway project between SR and Z39.50
with Danish, Irish and Portuguese partners.
The aim of the ION project, completed at the end of 1994, was
to develop interconnection between three computerised ILL
networks: LASER in the UK, PICA in the Netherlands and PEB
in France - in order to support and develop international
interlending and messaging services; and to demonstrate the
capabilities of the ILL and SR communication protocols.
IRIS is a document delivery service providing access via a
Z39.50 client developed by Fretwell-Downing to six Irish library
servers developed by Dynix, MDIS, BLCMP and Fretwell-
Downing. The user can select a group of library servers to for
distributed parallel searching . The Fretwell-Downing Z39.50
client (Oracle Libraries client) is evaluated above in
Appendix D.
The aim of the early NORDIC - SR-Net project was to establish
a network based on the SR protocol between library systems in
the 5 Nordic countries.
The objective of SOCKER is to demonstrate the viability of the
ISO SR protocol within various environments: CD-ROM
workstation, library system and network access point.
Appendix F.
Cataloguing and Resource Description DevelopmentsF.1 Introduction
If it is to be possible to find and retrieve EIOs, whether they are
documents, images or online catalogues, they must be
adequately described and their electronic location must also be
specified in such a way as to facilitate the retrieval of the EIO.
In other words, the gathering and encoding of meta-information
(information that is not found in the resource itself (e.g. source,
revisions) is required.
F.2 Cataloguing Requirements for the CATRIONA Model
The cataloguing requirements for the CATRIONA model
discussed below correspond to cataloguing requirements (a) to
(j) given at the end of each stage of the model description in
Appendix B. They were identified through an analysis of
Version 4 of the model.
Resource descriptions are required that cover all of the types of
electronic resource that users may want to retrieve. The record
must contain electronic location and access information. For
the CATRIONA model to function at a basic level, only minimal
details are needed (e.g. title, author, URL).
(d) Records must be of a type that can be searched and
retrieved using Z39.50.
This is a basic requirement for the CATRIONA model. Several
record formats can currently be transported using Z39.50
including USMARC, the record format central to the
CATRIONA proposals.
(f) Ability to store several alternative URLs in catalogue
record.
A record should be able to supply more than one URL for a
resource if more than one location exists.
This is required to in order to enable and control access to
commercial services.
(c) Ability to record that this is a special catalogue record only
to be retrieved/displayed if the search is a local one.
This is required in order to enable local librarians to manage
users' access to EIOs.
(e) Ability to catalogue and subject index other Z39.50 OPACs
and record Z39.50 information needed to set up client-server
dialogues.
Records should be able to describe Z39.50 OPACs, including
subject description and access information such as the address
of the host, number of the port, name of the database,
preferred record syntax and type of attributes supported.
(g) Ability to record Z39.50 information needed to set up
dialogues with non-MARC Z39.50 OPACs.
MARC records should be able to describe non-MARC Z39.50
OPACs and provide access information
(h) Records must be in a form that is compatible with
developing Internet standards
Standards adopted in the CATRIONA model should be
compatible with the emerging standards of Internet resource
description, such as those being developed by the IETF
(Internet Engineering Task Force) such as URCs.
(i) Ability to catalogue Internet indexing services such as
Archie, Veronica etc.
(Specific example of requirement (a))
(j) Ability to catalogue CD-ROM titles and record access
information for them.
(Specific example of requirement (a))
F.3 Current Implementations of Cataloguing Requirements
In this section, the way in which different cataloguing schemes
can accommodate the requirements of the CATRIONA model
is examined.
F.3.1 MARC Records
In 1992 OCLC began their first project in cataloguing Internet
resources. (Dillon et al, 1993). This involved 30 cataloguers
cataloguing 300 data resources, mainly documents. The results
of this project fed into Proposal 93-4 (MARBI, 1993a) for
cataloguing Internet resources. The most significant change
resulting from Proposal 93-4 was the addition of a new field
(856) "Electronic Location and Access". The field was accepted
in January 1993 and is now formally part of the USMARC
format. (See Note 1)
(a) Ability to catalogue any and every possible type of Internet
resource and service.
Yes - requirement met at a basic level. A wide range of Internet
resources can potentially be described using MARC and this is
to be tested in the current OCLC Project - Building a Catalog of
Internet Resources. (OCLC, 1994).
(d) Records must be in a form that can be searched and
retrieved using Z39.50.
Yes. The USMARC record syntax is a basic Z39.50 record
syntax.
(f) Ability to store several alternative URLs in catalogue
record.
Yes. This is possible as 856$u (USMARC) is a repeatable field.
"The field is repeated if more than one URL needs to be
recorded" (MARBI, 1993d).
(b) Ability to record username and password of commercial
services in suppressible MARC field.
It seems unlikely that this is met as it is probably specific to
CATRIONA.
(c) Ability to record that this is a special catalogue record
only
to be retrieved/displayed if the search is a local one.
It seems unlikely that this is met as it is probably specific to
CATRIONA.
(e) Ability to catalogue and subject index other Z39.50 OPACs
and record Z39.50 information needed to set up client-server
dialogues.
Probably specific to CATRIONA and unlikely to have been
considered.
(g) Ability to record Z39.50 information needed to set up
client-server dialogues with non-MARC Z39.50 OPACs.
Unlikely that this has been considered
(h) Records must be in a form that is compatible with
developing Internet standards
Compatibility is envisaged between MARC and Internet
resource descriptions or characteristics (URCs). "URCs will be
mapped algorithmically into and out of MARC records; to the
extent they are designed with this in mind, the Net and existing
libraries will work better together. MARC need not be the
syntactical wrapper for URCs, but the rules for encoding MARC
records (AACR-2) [Anglo-American Cataloguing Rules, second
edition) should inform those elements of the URC that are
similar" (Weibel, 1994)
(i) Ability to catalogue Internet indexing services such as
Archie, Veronica etc.
Yes, at a basic level of detail (e.g. title and URL in 856$u).
(j) Ability to catalogue CD-ROM titles of networked CD-ROMs
and record access information for them.
Extremely unlikely that this is met.
___________________________________________________________________________
Note 1: USMARC Format
Data elements for Electronic Resources
< > indicates change proposed in the most recent MARBI
Proposal 95-1 "Changes to Field 856 (Electronic Location and
Access) in the USMARC Bibliographic Format", December 2,
1994.
FIELD 856 (Additional explanation in italics)
Indicators
First Access method
0 Email
1 FTP
2 Remote login
< 3 Dial-up>
7 Method specified in subfield $2
Second Undefined
Undefined
Subfield codes
$a Host name (R) (R = Repeatable subfield)
$b
___________________________________________________________________________
F.3.2 IAFA Templates
IAFA (Internet Anonymous FTP Archives) templates were
designed to assist in searching through the vast amounts of
data stored at FTP archive sites. Although Archie was available
to do this, the metadata used by Archie is very sparse. IAFA
templates were developed in order to provide more information
to those searching archives, and to put the information into a
standard format, so that it could be machine-searchable.
(Weider, 1994)
IAFA templates are defined for a range of resources. These
include users, organisations, mailing lists, FTP archives, text
files, sound files, software packages and so on. (Deutsch et al,
1994).
(a) Ability to catalogue any and every possible type of Internet
resource and service.
Yes - requirement met at a basic level. A wide range of
resources and services can potentially be described in IAFA
templates:
USER, for individuals (also a cluster which appears in other
templates)
ORGANISATION, for organisations (also a cluster which
appears in other templates)
SITEINFO, for information about an FTP archive
LARCHIVE, for information about a logical archive within a
physical archive
mirror
'MIRROR', for mirror unit
'SERVICE', for mailing lists, video and sound feeds, and other
forms of interactive information sources
(Includes "online library catalogues ... interactive online
information services such as WAIS, gopher, prospero, WWW
or archie"
'DOCUMENT', for text files
'IMAGE', for image files
'SOUND', for sound files
'SOFTWARE', for software packages
'MAILARCHIVE', for mail archives
'USENET', for USENET archives
'FAQ', for Frequently Asked Questions (Deutsch et al, 1994)
No. IAFA templates are not yet registered as a record syntax
for transportation using Z39.50.
(f) Ability to store several alternative URLs in catalogue
record.
Seems possible. Example of document record gives two URLs
for different formats of the same text (ibid p34).
(b) Ability to record username and password of commercial
services in suppressible (MARC) field.
It seems unlikely that this is met as it is probably specific to
CATRIONA,
although in the SERVICE template there is an entry for
authentication information.
(ibid p9). See Note 3
(c) Ability to record that this is a special catalogue record only
to be retrieved/displayed if the search is a local one.
It seems unlikely that this is met as it is probably specific to
CATRIONA.
(e) Ability to catalogue and subject index other Z39.50 OPACs
and record Z39.50 information needed to set up client-server
dialogues.
Probably specific to CATRIONA and unlikely to have been
considered
(g) Ability to record Z39.50 information needed to set up client-
server dialogues with non-MARC Z39.50 OPACs.
Unlikely that this has been considered.
(h) Records must be in a form that is compatible with
developing Internet standards
Possible. IAFA templates and URCs (the current central
developing Internet standard of description) have similar data
elements and parseable formats and therefore interworking
may be possible. (Koster, 1994).
(i) Ability to catalogue Internet indexing services such as
Archie, Veronica etc.
Likely, at a basic level. The 'SERVICE' template covers "online
library catalogues, interactive online information services such
as WAIS, gopher, prospero, WWW or archie" (Deutsch et al,
1994 p9)
(j) Ability to catalogue CD-ROM titles and record access
information for them.
Extremely unlikely
___________________________________________________________________________
Note 3: Fields for the IAFA SERVICE template.
Template-Type: SERVICE
Title: Title of service.
URI: URI of service.
Admin-(USER*): Contact information of person
or group responsible for service
administration (administrative contact).
Owner-(ORGANIZATION*): Information on organization
responsible for this service.
Sponsoring-(ORGANIZATION*): Contact information for
the organization sponsoring this site.
Description: Free text description of service.
Authentication: Authentication information. Free text
field supplying login and
password information (if necessary) or other method for
authentication.
Registration: How to register for this service if
general access is not available.
Charging-Policy: Free text field describing any
charging mechanism in place.
Additionally, fee structure may
be included in this field.
Access-Policy: Policies and restrictions for using this
service.
Access-Times: Time ranges for mandatory or
preferred access of service.
Keywords: Keywords appropriate for describing
this service.
(Deutsch et al, 1994 p29-30)
___________________________________________________________________________
F.3.3 URC
The URI (Uniform Resource Identifiers) working group of the
IETF (Internet Engineering Task Force) has been developing
means of discovering, identifying and retrieving resources.
Resources are identified by a URN and retrieved by a URL.
"Describing the resource for purposes of discovery as well as
making the binding between a resource's name and its
location(s) is the role of the URC (Uniform Resource
Characteristic)". (Daniel, 1994b, p3)
a) Ability to catalogue any and every possible type of Internet
resource and service.
Requirement met at a much lower level than in MARC or IAFA
templates.
The emphasis of URCs at present seems to be on describing
retrievable resources. The examples given (ibid) are for
documents, i.e. text files. However, one of the requirements is
that "it must be possible to add new types of elements to the
URC without breaking previous applications" (Daniel, 1994b
p16)
(d) Records must be in a form that can be searched and
retrieved using Z39.50.
No. URCs are not yet registered as a record syntax for
transportation using Z39.50.
(f) Ability to store several alternative URLs in catalogue
record.
Yes. An example is given with two URLs. (Mealling, 1994).
(b) Ability to record username and password of commercial
services in suppressible (MARC) field.
No.
(c) Ability to record that this is a special catalogue record only
to be retrieved/displayed if the search is a local one.
No
(e) Ability to catalogue and subject index other Z39.50 OPACs
and record Z39.50 information needed to set up client-server
dialogues.
No
(g) Ability to record Z39.50 information needed to set up client-
server dialogues with non-MARC Z39.50 OPACs.
No
(h) Records must be in a form that is compatible with
developing Internet standards
Yes. The URC is the current central developing Internet
standard of description.
(i) Ability to catalogue Internet indexing services such as
Archie, Veronica etc.
Likely at a basic level.
(j) Ability to catalogue CD-ROM titles and record access
information for them.
No
___________________________________________________________________________
Note 4: URC Structure
Attributes: Values:
URN Uniform Resource Name
URL Uniform Resource Locator
LIFN Location Independent File Name
(to be developed)
Author Author of the Item (there are no requirements
as to how a name should be written)
TTL Time To Live (measured in seconds, this sets a time limit
on the attribute/value pair preceding it and is meant as
a caching aid)
Abstract A short abstract about the item. Any characters
are admissible (Mealling, 1994).
Note 5:
Examples of URC attribute-value pairs
Attribute: Value
URN: IANA:626:oit.5647
URL: http:www.gatech.edu/ietf/urc.encoding.html
LIFN: md5:2039874029834059283709475029387405
928374095827394875
Author: Michael Mealling
TTL: 86400
Abstract: This document explores the various flight
patterns and speeds of unladen African and
European swallows. A companion document
concerning the relative velocities of swallows
laden with coconuts is available. (Mealling, 1994).
___________________________________________________________________________
F.3.4 TEI Header
The TEI (Text Encoding Initiative) has developed guidelines for
the preparation and exchange of electronic texts mainly for
research in the humanities. Meta-information relating to the
encoded text is contained in the TEI header which can be in a
different file from the main text (as an Independent header).
The TEI header consists of four elements: File Description;
Encoding Description; Profile Description and Revision
Description. The only mandatory element is File Description,
which provides a bibliographic description very similar to
standard descriptions: ACCR-2, ISBD (International Standard
Bibliographic Description) (G) and USMARC. (Giordano,
1994a) (See Note 6 for key elements in FileDesc).
(a) Ability to catalogue any and every possible type of Internet
resource and service.
No. TEI has been developed for encoding and describing
electronic texts.
(d) Records must be in a form that can be searched and
retrieved using Z39.50.
No. TEI headers are not yet registered as a record syntax for
transportation using Z39.50.
(f) Ability to store several alternative URLs in catalogue
record.
No. There is place at present allocated for a URL in a TEI
header. (Giordano, 1994b)
(b) Ability to record username and password of commercial
services in suppressible (MARC) field.
No.
(c) Ability to record that this is a special catalogue record
only
to be retrieved/displayed if the search is a local one.
No.
(e) Ability to catalogue and subject index other Z39.50 OPACs
and record Z39.50 information needed to set up client-server
dialogues.
No.
(g) Ability to record Z39.50 information needed to set up
client-
server dialogues with non-MARC Z39.50 OPACs.
No.
(h) Records must be in a form that is compatible with
developing Internet standards.
Mapping with MARC is possible with some of the elements,
although the TEI header contains information which does not
have equivalent MARC fields. There has also been recent
discussion of a TEI-URC relationship (Herr-Hoyman, 1994).
(i) Ability to catalogue Internet indexing services such as
Archie, Veronica etc.
No.
(j) Ability to catalogue CD-ROM titles and record access
information for them.
No.
________________________________________________________________________
Note 6: TEI Header Elements
The following is a list of a few of the key components of the
TEI header selected from an extensive list of possible
elements.
*
________________________________________________________________________
F.3.5 Other Developments: GILS
GILS (Government Information Locator Service) is a US-based
project promoted by the US Information Infrastructure Task
Force and the Office of Management and Budget. The purpose
of GILS is to identify and describe government information
resources and provide help for users to obtain the information.
The public will be able to use GILS either directly through the
Internet or indirectly through various intermediaries including
public libraries. GILS will use the Z39.50 standard in order to
make information as widely available as possible. Because
GILS will be a decentralised service, there will be a wide range
of implementations and the need to agree on a reference base
of stable standards including Z39.50 to achieve interoperability
was recognised. In addition, a GILS Profile has been developed
which gives a complete specification of GILS as it makes use
of ANSI Z39.50, but also specifies the features of GILS that are
outwith the scope of ANSI Z39.50. In the GILS draft profile it is
planned that GILS records will be available in three Z39.50
record
syntaxes: USMARC, GRS and SUTRS.
___________________________________________________________________________
Note 7: GILS Core Elements
Title: (USMARC Tag 245$a)
Control Identifier: (USMARC Tag 001)
Abstract: (USMARC Tag 520)
Purpose: (USMARC Tag 500)
Originator: (USMARC Tag 710$a)
Access Constraints: (USMARC Tag 506)
Use Constraints: (USMARC Tag 540)
Availability: (USMARC Tag 037$b)
Resource Description: Optional sub-element (USMARC Tag 037$f)
Order Process: Mandatory sub-element (how to obtain the
information resource from this distributor) (USMARC Tag037$c)
Technical Prerequisites: Optional sub-element (USMARC Tag 538)
Available Time Period: Optional sub-element
Available Linkage: This optional sub-element occurs no more
than once per Availability element. It provides the information
needed to contact an automated system made
available by this distributor, expressed in
a form that can be interpreted by a computer
(i.e., URI).
Available linkages are appropriate to
reference other locators, facilitate
electronic delivery of off-the-shelf
information products, or guide the
user to data systems that support analysis
and synthesis of information. (USMARC Tag 856$u)
Available Linkage Type: This optional sub-element occurs if there
is an Available Linkage described. It provides the data content
type (i.e., MIME) for the referenced URI.
(USMARC Tag 856 first indicator/856$2)
Point of Contact for further
information: This mandatory element occurs once per
locator record. (USMARC Tag 856$m for
electronic resources
(Christian, 1994)
Return to Contents Page
Appendix G.
CATRIONA - A Case Study: Napier University LibraryG.1. Background
Napier University is a new University, formerly Napier
Polytechnic of Edinburgh, spread across several campuses.
Napier University Library (NUL) has full-service library facilities
at three major campuses, with a fourth and largest campus
currently under development. The student population is just
over 9000, and there are 1000 staff. The University has an
Ethernet LAN which is scheduled for upgrading and connection
to the Edinburgh Metropolitan Area Network (MAN). Access to
the Library automated catalogue is via the LAN; there are no
public terminals or PCs connected directly to the Library
system.
G.2 Current Situation
NUL believes an integrated, one-stop, public-access catalogue
to be the best tool to meet Strategic Plan objectives on self-
service and a mixture of information formats. The catalogue
comprises UKMARC standard records, using Anglo-American
Cataloguing Rules, 2nd edition, 1988 revision, 2nd level of
description, for most of NULs stock of physical formats,
including monographs, serials, sheet music, music scores,
video recordings, audio Compact Discs, and stand-alone CD-
ROMs. In addition, attempts have been made to catalogue
CD-ROMs and Windows/DOS packages which are available
only via the University LAN, and therefore have no physical
manifestation as such. The resulting integration of access to
these resources via the catalogue is illustrated by Screen 1.
Screen 1: Integrating resources concerning the Guardian
newspaper: title search
Alignment of title numbers indicates whether the material is
available at the local campus library.
Your search: The Guardian
TITLE/AUTHOR EDITION
DATE
1. The bedside Guardian 27
Webb, W. L.
1978
2. The Guardian
3. Guardian: biography of a newspaper
Ayerst, David
1971
4. The Guardian on CD-ROM
Guardian Newspapers Ltd
5. The Guardian on CD-ROM including The Observer
Guardian Newspapers Ltd
6. Raising finance: the Guardian guide for the sma
Woodcock, Clive 2nd ed
1985
- - - - 6 titles, End of List - -
- -
Title 2 is the printed newspaper.
Title 3 is a book about the newspaper.
Title 4 is the previous title of the newspaper archive on CD-
ROM
Title 5 is the current title of the newspaper archive on CD-
ROM.
Title 6 is a book written by staff of the newspaper.
Screen 2: Card format display of selected title
DYNIX BIB # 104199
TITLE The Guardian on CD-ROM including The
Observer
The Guardian
The Observer
The Guardian on CD-ROM
NAME(S) 1) Guardian Newspapers Ltd
PUBLISHER [London] ; Chadwyck-Healey ;
SIZE, ETC. computer laser optical disk
NOTES Issues cumulate within the year
Continues: The Guardian on CD-ROM, 1994
'Guardian Newspapers Limited'
SUBJECT 1) 072.1
FORMERLY The Guardian on CD-ROM
FREQUENCY Quarterly
- - - - End of Title Info - - -
-
Screen 3: Dummy holdings record instructing user how to
access the resource
Hyphens are used to fill the mandatory Shelfmark field in each
holdings record
Author Guardian Newspapers Ltd
Title The Guardian on CD-ROM including The
Observer Holds: 0
# CALL # STATUS BARC
ODE LIB
1. Circulation desk On shelf 002185245
S
- Manual
2. Circulation desk On shelf 002071171
M
- Manual
3. Circulation desk On shelf 002071510
C
- Manual
4. See Network menu CD-ROM Network -144891
CN
-
Holding 4 refers to the CD-ROM mounted on the LAN. It would
be misleading for users, and cause operational problems, to
locate the CD-ROM at any of the campus libraries; NUL has
therefore set up an extra library, coded CN, to indicate where
the material is held. A barcode is mandatory for each holding;
in this case a dummy or negative barcode is used because the
resource has no physical manifestation.
G.3 Developing the Current Approach
NUL anticipates a rapid increase in the number of electronic
information resources it provides. In addition to the non-print
resources already mentioned, NUL expects to have to stock or
provide access to Windows/Multi-media teaching and reference
packages, CD-i discs, and CD-Video discs, all of which could
be mounted on the LAN and connected directly to an
appropriate PC workstation. Many print-based reference
services are being replaced by electronic versions distributed
by both local carriers such as CD-ROM or DOS file and remote
dial-up systems similar to BIDS. NUL also wishes to offer
seamless access to Internet resources and services, for
researchers and students, which have been deemed of interest
and relevance. NUL does not believe that menu-driven,
hierarchical access tools will be able to cope with more than a
few hundred electronic resources, whether on the LAN, WAN or
MAN, and in any case requires that access and navigation tools
for these resources to be integrated with those for traditional
resources held as physical items in stock so that self-service is
encouraged, and Library staff are not required to mediate
between identification of a resource and access to it.
Screen 4: Potential development of the existing approach: title
search
Your search: The Guardian
TITLE/AUTHOR EDITION
DATE
1. The bedside Guardian 27
Webb, W. L.
1978
2. Editing the Guardian: an interactive experienc
Compact Eyes Productions
3. The Guardian
4. Guardian: biography of a newspaper
Ayerst, David
1971
5. The Guardian on CD-ROM
Guardian Newspapers Ltd
6. The Guardian on CD-ROM including The Observer
Guardian Newspapers Ltd
7. The Guardian online
Guardian Newspapers Ltd
8. The photo-journalist: an archive of famous pho
Press Photo Archives plc
9. Press reviews: an Internet service
Reviews in the News, Inc.
10. Raising finance: the Guardian guide for the sma
Woodcock, Clive 2nd ed
1985
- - - -10 titles, End of List - -
- -
Screen 5: Potential development of the existing approach: title
detail screen
DYNIX BIB # 999999
TITLE The Guardian online
The Guardian
The Observer
NAME(S) 1) Guardian Newspapers Ltd
PUBLISHER [London] ; Guardian Online Ltd ;
NOTES Online access via JANET
Access available to registered users of the
University Network
Mosaic:
http://gol.news.co.uk:6560/www/gol.html
SUBJECT 1) 072.1
- - - - End of Title Info - - -
-
G.4 The Preferred Approach
NUL anticipates a major potential problem in the sheer quantity
of networked resources that may require cataloguing or some
other access and navigation system. NUL would seek to
identify and provided integrated access to resources likely to be
of a high interest and relevance to its users, but would also
wish for extended access to other similar resources via other
libraries. This idea already applies to physical resources: if a
user cannot find what they want in NULs catalogue, they are
encouraged to search the catalogues of other libraries in the
vicinity, for which there are reciprocal user agreements. NUL
sees the use of Z39.50 protocols as an essential development
to allow extended searching beyond NUL itself. NUL is not
concerned with the source of the catalogue records: they may
be held in the local database, or the catalogue of another local
library, or by the resource provider. The resource has no
physical manifestation and its location is networked information
space; concepts of shelfmark, library, holdings, circulation
status, and other storage parameters, are replaced by filename,
URL, usage license, browser, viewer, and other access
parameters.
Appendix H.
Reports of Meetings on Z39.50 DevelopmentsH.1 Report of a Joint Meeting on Z39.50
The Library Association Information Technology Group (Central
Scotland) and the Cataloguing & Indexing Group in Scotland
held at joint meeting on Z39.50, with a demonstration of the
SALSER database, at Glasgow University Library during the
afternoon of 18 January 1995.
H.2 Z39.50 Pre-Implementors Group Meeting
Aston University Library, 18 October 1994
The meeting began with a demonstration of the SLS Libertas
6.3 Z39.50 client. The facility to select groups of servers for
distributed searching was effectively demonstrated. (See
Appendix D for a list of client features).
H.3 Aston University Seminar: Z39.50 UK Wide: What's Going
On
?
A seminar at Aston University organised on behalf of the
Z39.50 Pre-Implementor's Group
The seminar, which was attended by systems suppliers,
software vendors and LIS professionals, was designed to
introduce the latest developments in Z39.50 implementation to
the LIS community. A keynote address on Z39.50 strategy in
the UK was given by Lorcan Dempsey of UKOLN. After
outlining the Z39.50 protocol, he emphasised the current
obstacles to progress in the UK: the immaturity of current
distributed library systems; the lack of significant UK resources
accessible via Z39.50 and lack of resource-sharing; and the
need for experience of how the servers will work. This was
followed by descriptions of a wide range of UK and European
Z39.50/SR projects. The CATRIONA project was mentioned
and was enquired about in during the question time which
followed.
Aston University Library, 24 January 1995H.4 Meeting of the SCOLAR Technical Sub-Committee on
Z39.50
The meeting of the SCOLAR Technical Sub-Committee which
was held on 5 December, 1994, was addressed by Ameritch
representatives Tom MacDonald, Maribeth Ward and John
Kolman. After an outline of Z39.50, the Ameritech client
WinPAC which was originally developed by NOTIS, was
introduced. (WinPAC is one of the clients that was examined
during the CATRIONA project. See Appendix D). Future plans
for WinPAC which include expanded access were discussed,
as well as the need for developments on the Z39.50 server
front such as different indexing techniques and more
sophisticated software and hardware.
Appendix I.
BUBL Subject Tree Project and Relationship to CATRIONAI.1 BUBL Subject Tree Initiative
The BUBL Subject Tree Initiative was begun in the final quarter
of 1993 and is now a significant part of the BUBL Service with
over 5000 links by June 1994. BUBL was the first national UK
service to offer subject-based access to Internet services and
resources and is still the only UK national service whose
subject tree covers all main subjects. It is also unusual,
perhaps unique, in that it provides a composite tree covering
both Gopher-based resources and World-Wide Web resources.
The tree may be accessed in a number of ways:
http://www.bubl.bath.ac.uk/BUBL/Tree.html
and
http://www.bubl.bath.ac.uk/BUBL/Treealphabet.html
Name=BUBL Information Service
Type=1
Port=7070
Path=
Host=ukoln.bath.ac.uk
http://www.bubl.bath.ac.uk
I.2 The BUBL Subject Tree and CATRIONA
BUBL regards the Subject Tree as a useful interim measure. In
the long term the only solution resource discovery on the
Internet is likely to be based on local cataloguing and national
and international co-operative cataloguing. It is also likely that
the voluntary effort on which the BUBL Subject Tree is based
will have problems in scaling up to cope with the expansion of
electronic information. These are some of the reasons behind
the CATRIONA project which arose from the BUBL Subject
Tree Initiative. If CATRIONA Phase II goes ahead it may
ultimately replace subject trees by replacing them with an
improved service, but this is not likely to happen for some time
to come.
I.3 Further Information
Further information on the CATRIONA feasibility study, on
other related information and projects, and on the BUBL
subject tree will be found on the BUBL Information Service at
URLs:
http://www.bubl.bath.ac.uk/BUBL/maincatriona.html
gopher://bubl.bath.ac.uk:7070/11/Link/Catriona
http://www.bubl.bath.ac.uk/BUBL/Treealphabet.html
gopher://www.bubl.bath.ac.uk:7070/11/Link/TreeReturn to Contents Page
Appendix J.
The organisations and individuals who participated in the
CATRIONA Project are listed below. Participation was at a wide
range of levels.
List of Project Participants, Advisors and Correspondents
DYNIX Library Systems (UK) Ltd.
Fretwell-Downing Informatics Ltd
SLS Information Systems
SIRSI Limited
MDIS (McDonnell Information Systems)
SCG (Specialist Computer Group)
SilverPlatter Information, Ltd
OCLC (Online Computer Library Center)
RLG (Research Libraries Group)
CURL (Consortium of University Research Libraries)
SCURL (Scottish Confederation of University and Research
Libraries)
CIGS (Cataloguing and Indexing Group Scotland)
BLCMP Library Services Ltd
Lorcan Dempsey, UKOLN: the UK Office for Library and
Information Networking
Hunter Monroe, Eric Lease Morgan (Alex Project - See
Appendix E)
The British Library
Glasgow University Library
Stirling University Library
Abertay University Library
Paisley University Library
Edinburgh University Library
Glasgow Caledonian University Library
Strathclyde University Library
National Library of Scotland
Napier University Library
Heriot Watt University Library
Aberdeen University Library
Appendix K.
This Workshop was one of two presented during the 7th annual
OPAC Forum, organised by Cataloguing & Indexing Group in
Scotland (CIGS) and held in the Library of the University of
Abertay Dundee on 8 November 1994. The Workshop leader
was Gordon Dunsire, Information Systems Librarian at Napier
University, Edinburgh; the notes on which this report is based
were taken by Lynn Corrigan, Assistant Librarian (Information
Systems), also of Napier University.
Workshop on Cataloguing Electronic ResourcesPresentation
The presentation gave an overview of some of the issues
concerning resources, records, and the catalogue.
Discussion
The fundamental assumption, that this meta-catalogue is a
good thing, was queried. Many users are only interested in
local resources, and the old-fashioned OPAC might be
sufficient for their needs. It was suggested that local really
meant available, and in that sense all network EIOs could be
considered local if it was possible to deliver them to the OPAC
or workstation screen.
Return to Contents Page