Staff Papers

Directory services

A paper presented by Judith Pearce, National Library of Australia, at the Standards Australia/Standards New Zealand Committee IT/19 seminar held in Melbourne, Australia, on Monday 21 February, 2000.


Introduction

In an unusual venture, IT/19 is taking the lead role in the revision of an existing ISO standard. The standard is ISO 2146 Documentation -- Directories of libraries, archives, information and documentation centre, and their databases. Its purpose is to assist in compiling and publishing directories of organisations operating in the library and information domain.

ISO 2146 was first published in 1972 by a former subcommittee of ISO Technical Committee 46 (the Subcommittee on Documentation). It was revised in 1988 and responsibility passed to Subcommittee 9.

Subcommittee 9 deals with the presentation, identification and description of documents. ISO 2146 sat quite well with the other responsibilities of this subcommittee in pre-Web days when the primary means of publishing directory information was as a printed document. However, it was not a mainstream interest of the subcommittee. Neither was the standard itself taken up in any widespread way by implementers. When it came up for its ten year review in 1998 Subcommittee 9 was therefore in the process of deprecating it when IT/19 intervened.

ISO TC46 Sub-committee 4, Working group 7 will lead the revision of ISO 2146. Janifer Gatenby chairs this working group. My role in the project will be to coordinate committee drafts.

Why TC46/ SC4 WG7

TC46 Sub-committee 4 deals with computer applications in information and documentation. Working Group 7 is responsible for data elements. It has already produced data element directories for a range of library applications under the umbrella name of ISO 8459.

A data element directory defines the data elements needed by a computer application to support a specific set of user scenarios. The definitions are independent of any specific protocol or record syntax. The directory acts as a checklist for designers building applications and ensures a conformity of approach.

Parts 1-5 of ISO 8459 cover data elements for interloan, acquisitions, information retrieval, circulation and cataloguing and metadata applications. A current task of the working group is to consolidate all the parts into a unified data dictionary standard.

The ISO 2146 Project thus fits well under Working Group 7. The aim of the project is to transform ISO 2146 from a standard for the collection of data for print publishing to a generic data element directory that can be used by a range of directory applications. This will extend the coverage of the ISO 8459 family of data element directories to include directory data elements.

The need for directory data

Directory data informs many of the processes needed for information delivery in a distributed network environment.

In addition, service providers need to build and maintain directories of stakeholders and to access directory information to facilitate the service delivery process, for example when they need to negotiate with another service provider to obtain a copy of an information resource on behalf of a user.

Inadequacies of ISO 2146

In its current state ISO 2146 is not suitable as a data element directory for applications addressing these user scenarios. Data elements are tied to a specific syntax. There is no underlying data model. Relationships between organisations and organisational units are expressed in a linear fashion for print delivery. Issues such as the difference between collection, database and service and how these entities relate to each other remain unresolved.

For these reasons alone, the standard needs to be completely rewritten using the framework being developed for the consolidation of ISO 8459. In addition, substantial revisions are needed to the data elements themselves to take into account new usage scenarios, to reflect advances in technology since 1988 and to acknowledge related standards.

Another strategy for achieving the same objectives would have been to propose a new work package and to develop a completely new standard. By using ISO 2146 as a starting point, we are hoping to fast-track the process. In addition, the new version will build on the intellectual effort and experience with directory data that went into development of the existing standard.

Dublin Core resource types

To understand the range of data elements that fall within scope of the ISO 2146 Project, it is useful to look at the list of proposed Dublin Core resource types. You will recall that Dublin Core is intended for the simple description of electronic resources and that the resource type element is used to categorise the nature of the resource. The values proposed for this element are:

Dataset, image, interactive resource, model, software, sound and text can all be seen as examples of information objects: objects that are of primary interest for their information content.

Physical object is in the list primarily to meet the needs of the museum community. A digital image of a cultural artefact has as its source a physical object that also needs to be described.

Collection, event, party, place and service are all examples of objects that can be seen as directory objects and that fall within the scope of the ISO 2146 Project. Libraries, archives and information and documentation centres are parties. They own or manage collections, they deliver services based on these collections and they engage in activities that may be described as events.

Directory objects

There has been some debate about the correctness of defining directory objects as Dublin Core resource types, but they are useful in a number of contexts. For example, they allow a website manager to use the same metadata standard across the whole site while distinguishing between agency publications, service entry points, conference pages and personal home pages. Someone building an events database, however, or an interlibrary resource sharing directory or a name authority file would find it difficult to apply Dublin Core for these specific purposes without substantial extension.

The Dublin Core Metadata Initiative recognises this and acknowledges the need for a Web infrastructure that “supports the coexistence of complementary, independently maintained metadata packages”. The same object may be described by different packages for different purposes; and packages may be linked to enable navigation from one object type to another.

A range of standards has been developed by various communities that define “metadata packages” or “data schemas” for each of these directory object types. One of the terms of reference for the ISO 2146 Project will be to ensure these standards are taken into account when developing the data element directory.

In this paper, I will introduce you to these standards and the user scenarios they address and I’ll also look at three important works in progress. But first I will briefly summarise the range of protocols and record syntaxes that these standards use and that are applicable to directory data.

Protocols are the messages exchanged between a client and server in an application, for example, to enable search, retrieval and update of data. Record syntaxes are ways in which data elements making up a data schema are structured for exchange between applications. While the new version of ISO 2146 will be protocol and syntax independent it will be developed with an awareness of the ways in which directory data is likely to be queried, updated and exchanged.

Directory protocols

A number of protocols exist that have their origins in the need to query, update and exchange directory data. Many of these protocols have been extended for use with other object types. Similarly, directory data may be queried through protocols which have not been developed specifically for handling directory objects. The key protocols of interest to the library and information community are:

X.500 is a standard for distributed directory services. The standard provides both an information model for the X.500 database and also the protocol used to query and update the database. X.500 can be used for different types of directories but its most notable implementation is as a global White Pages service.

LDAP (the Lightweight Directory Access Protocol) is a lightweight version of X.500 often implemented as a stand-alone database without access to wider directory services. X.500 and LDAP share the same information model and can usually be treated together in discussions of directory data standards.

WHOIS++ is a simple Internet protocol originally developed to locate networks, domains and hosts on the Internet and to record contact information for them. The protocol has now been extended to enable query of distributed databases containing records of any structure. Record schemas are defined through a series of templates.

In the library and information world the most notable implementation of WHOIS++ is the ROADS software. ROADS stands for Resource Organisation And Discovery in Subject-based services. The ROADS software was developed through a project of the UK elib (electronic libraries) programme and is now in use by a number of subject gateways in the United Kingdom and elsewhere. Examples are OMNI (health and medicine) and SOSIG (social sciences).

Z39.50 is a functionally-rich search and retrieval protocol that can be used to query and update distributed databases containing records of any structure. It has been widely implemented in the library and information community for resource discovery and collection management purposes. Record schemas are defined outside the protocol through profiles such as the Bath and CIMI profiles.

HTTP is the ubiquitous query syntax of the World Wide Web. Its primary use is to get static HTML documents from distributed servers using a known resource identifier. Through gateways with other protocols it can also be used to query databases of structured data.

XML query language is a current initiative of the World Wide Web. The aim of this initiative is to develop a query language designed specifically for XML documents in the same way that SQL is a query language for relational tables.

Directory data is likely to be accessed through a range of protocols in the short to medium term. Particularly important will be the implementation of Z39.50/X.500, Z39.50/WHOIS++ and WHOIS++/X.500 gateways to query distributed databases that use protocols not supported by the client application. An application may even need to use different protocols within one business transaction. For example, the Z39.50 Implementers' Group has agreed on a mechanism for limiting a search of a union catalogue to items in, say, Melbourne libraries; and for limiting subsequent record displays to the holdings of these libraries only. But how does the application know which libraries are in Melbourne?

Directory record syntaxes

For applications that are not based on any specific data schema, the only way to exchange directory data may be as files of records in ASCII format. This depends on the receiving application having some knowledge of the data schema.

WHOIS++ implementations exchange records in ASCII format using simple attribute / value pairs defined by ROADS templates and special characters to indicate the start and end of data elements and records.

MARC is used as the record syntax for data in MARC formats such as the USMARC format for community information.

XML can support a more complex data model than ASCII and is becoming the record syntax of choice for many applications.

Z39.50 implementations can negotiate with each other to exchange records in a variety of syntaxes including MARC, HTML, SGML, XML SUTRS and GRS-1. SUTRS is a syntax for the retrieval of simple unstructured text. GRS-1 is a generic record structure that uses tag paths in combination with a known data schema to enable the content of a record to be parsed.

LDIF is the LDAP data interchange format. Its primary use is to exchange data between an LDAP directory and another application without using LDAP protocols. It is also starting to be used as the means of exchanging data between an X.500 directory and another application.

Directory Standards

Party

The DC party resource type is defined as “a person, organization, cultural group, or institution”. Information commonly needing to be stored about parties include names, physical and electronic addresses and relationships between parties.

The X.500 information model provides a comprehensive solution for any “white-page” directory service, with object classes defined for:

Vcard is a standard developed by the Versit Consortium. It uses the X.500 person object as the basis for an electronic business card. Additional attributes defined by Vcard include logos, photos and digital sound chips.

The Dublin Core Metadata Initiative is planning to use the Vcard data schema as a means of linking “agent” details to a separate metadata package detailing information about the agent.

Similarly, the ROADS Project has defined what are termed “cluster templates” for individuals and organisations. Cluster templates enable data elements such as name, address, telephone number and e-mail address to be recorded once in a ROADS application then referred to in other templates via a unique handle.

In the library and information domain two important standards are in use for name authority control – Library of Congress Name Authorities, which use AACR2 as the basis for the form of names; and ISAAR(CPF), an authority record for corporate bodies, persons and families used widely by the archive community.

Resource discovery services tend to be collection-based but a new model is emerging that uses name authority records to link resources that may be distributed in a variety of collections. This is a model familiar to the archive comunity and implemented in services such as the Australian Science Archive Project's Bright Sparks site.

The newly developed Z39.50 Thesaurus Profile provides an abstract record structure and access points for searching name authority records. This will help in the deployment of name authorities in a distributed environment.

Gavan McCarthy from the Australian Science Archives Project has been working with an international group on a task to produce an XML DTD based on ISAAR(CPF). This will provide an XML record syntax for archival authority records that can be exploited by applications using the XML query language or proprietary protocols as well as Z39.50.

ISAAR(CPF) diverges slightly from AACR2, an issue when searching across library and archival resources that might need to be addressed through unique resource identifiers for people and organisations.

Place

The DC place resource type is defined as “a geographic area”. Like party, place can be a directory object in its own right and also a property of another object. The X.500 information model includes a Country object class as the highest level in its hierarchy as well as attributes to describe the location of other directory objects. A range of standards exist for specifying location. These include LCSH geographic subdivisions, the USMARC Code List for Countries, ISO 3166 Country Codes, national and international authorities for place names, postcodes and zipcodes and geo-positioning coordinates.

Collection

The DC collection resource type is defined as “an aggregation of items”.

Resource description standards such as Dublin Core and the MARC bibliographic format can be used to describe collections when it is appropriate to have a single high-level entry point for the collection in a metadata repository or catalogue.

Specialist formats exist for the description of archival collections. These include the MARC format for archival and manuscript control, ISAD(G) (the General International Standard Archival Description) and EAD (the Encoded Archival Description), an XML Document Type Definition based on ISAD(G). A Z39.50 profile based on ISAD(G) has been developed and tested.

None of these standards deal with the data elements needed to describe whole library collections or substantial formed or specialist collections that are managed and searched in their own right. Later I will describe some work in progress in this area through the eLib Programme’s Collection Level Description Project.

The Conspectus mechanism developed by the Research Libraries Group in 1979 defines data elements for describing collection strengths and weaknesses. Several library directory services are exploiting this data for resource sharing and discovery.

SCURL (The Scottish Confederation of University and Research Libraries) has implemented a Research Collections Online database with Conspectus details for a number of Scottish university and research libraries. Their intention is to use this database to enable subject-based clump searching of the libraries’ catalogues using the Z39.50 protocol. There are issues still to be resolved re updating the database, adding new collections and discovering the Z39.50 services and facilities supported by specific servers.

The Australian Libraries Gateway has also implemented data elements for recording and searching on library collection strengths. It will be important for projects like SCURL and ALG to agree on data standards and protocols in this area if subject-based clump searching outside national or regional boundaries is to occur.

Service

The DC service resource type is defined as “a system that provides one or more functions of value to the end-user. Examples include: a photocopying service, a banking service, an authentication service, interlibrary loans, a Z39.50 or Web server.”

The object classes defined within the X.500 standard do not cover “yellow-page” services. Schemas to extend the X.500 information model to cover service descriptions are being developed within specific communities and through specific profiling activities. Later I will describe some work in progress in this area through the Collection Level Description and PRIDE Projects.

There are several existing data standards that describe services.

The Z39.50 Explain Service is a means of describing specific Z39.50 services and facilities supported by a Z39.50 server. Explain has not been widely implemented and a number of non-standard databases now exist that contain connection details, database names and a description of attribute sets and record syntaxes supported.

ROADS has a service template which contains data elements for access and charging policies and access times.

ROADS has also implemented a mechanism enabling a ROADS client to decide which subject gateway services are likely to contain information meeting a user’s needs. A centroid is essentially a simple inverted index of the information in the ROADS database. For each of the types of record that have been used and each of the attributes within those records, the centroid contains a list of all the index terms that have been used. Centroids are in effect data elements that a directory service might store for a given database-type service. This means that they could be exploited by other protocols such as Z39.50.

Standards for the description of offline services such as library opening hours or interlibrary loan services are still very undeveloped. There are some data elements in the Z39.50 holdings schema related to lending policy but these are specified at the collection item level. Later I’ll talk about what the Interlibrary Loan Profile Implementer's Group (IPIG) and the PRIDE Project are doing in this area at the service level.

Serving a quite different market is the quite recently developed USMARC format for community information. This standard has a record type for program or service which it defines as data relating to activity that carries out the purpose of an organisation or group (e.g., driver eduction, blood bank)”. In the United Kingdom, public libraries are exploring an architecture that will use Z39.50 to link distributed databases of community information based on this format.

Event

The DC event resource type is defined as “a non-persistent, time-based occurence. Metadata for an event provides descriptive information that is the basis for discovery of the purpose, location, duration, responsible agents, and links to related events and resources. The resource of type event may not be retrievable if the described instantiation has expired or is yet to occur. Examples - exhibition, web-cast, conference, workshop, open-day, performance, battle, trial, wedding, tea-party, conflagration.”

Project is not mentioned in the Dublin Core definition. I am assuming a project is also an example of event. A project is time-based and has purpose, location, duration, responsible agents and links to related projects and resources.

Most of the standards work in this area has centred on the development of protocols and data schemas by the Internet Engineering Taskforce (IETF) to support distributed calendaring and scheduling services. This has resulted in a core object specification known as iCalendar. Data elements for events are not well developed in the iCalendar specification but a lot of work has been done by the IETF working group to resolve international date/time issues.

The USMARC format for community information has a record type for event which it defines as “a record in which the data relate to a scheduled happening (e.g., lecture, regularly scheduled meeting).”

ROADS also has an events template.

Work in progress

Three three important projects are starting to fill the gaps identified in this audit of existing directory data standards:

The Collection Level Description initiative focuses on the needs of the higher and further education communities to discover the existence of collections meeting their needs. The IPIG ILL Directory initiative focuses on the needs of the interlibrary loan and document delivery communities. The PRIDE X.500 schema initiative addresses both these areas with specific emphasis on the development of an access management architecture.

Other communities are also monitoring this work. For example the Consortium for the Interchange of Museum Information (CIMI) is looking at the need for a directory of organisations conformant to the CIMI Profile. There is also a strong interest by the museum and archive communities in standards facilitating the development of gateways such as the Australian Library Gateway that are based on structured databases of information rather than static HTML pages.

Collection Level Description

The Collection Level Description was developed by an eLib working group to address the need of the eLib Hybrid Libraries and Large Scale Resource Discovery (Clumps) Projects to describe collections of resources.

The outcome was a collection level description with 23 data elements, 12 taken from Dublin Core. Data elements are included for describing collections themselves and services providing access to collections. Data elements for describing collection owners are left to other standards such as X.500 and WHOIS++.

Like Dublin Core the collection level description supports a simple descriptive model enabling resource discovery applications to address two of the user scenarios I outlined earlier in this paper:

Data elements for recording access and charging policies are not yet sufficiently developed in the Collection Level Description to enable unmediated access to services and service delivery options.

The collection level description is currently being tested through an experimental ROADS collection template. The work of this group has informed development of the PRIDE schema and is also being tested and further developed through the RSLP Collection Description Project. RSLP stands for Research Support Libraries Programme. Deliverables of this project will be an RSLP collection description schema, a simple web-based collection description editor, data entry guidelines and a prototype search service.

One of the early outputs of the RSLP Collection Description Project has been an analytical model of collections and their catalogues which builds on the earlier eLib work. Central to this work, authored by Michael Healey, is the idea that a collection can be not just an aggregation of resource items – for example, information objects in a library or physical objects in a museum - but that it can also be an aggregation of resource item descriptions – a catalogue, an index or a subject gateway. This enables information about services needed to access a collection to be separated from information about the collection itself.

IPIG ILL Directory

IPIG is the Interlibrary Loan Protocol Implementors’ Group. Several years ago IPIG drafted a set of data elements for an ILL Directory. These data elements are now being revised in the form of an XML schema. A web form will be developed to enable IPIG members to update information about their ILL services. Members will post the information on their web sites where it can be harvested for access via a central repository or for use by specific applications. A standard query protocol for this initiative is still undefined.

Data elements are defined in the 1997 draft for the organisation and its interlibrary loan and document delivery processing units, ILL symbols and codes, membership of union catalogues, ILL messaging systems used, types of material supplied, delivery methods, physical shipping address and shipping instructions, electronic addresses, ILL protocol information, service types, level of service provided, clientele type, international lending supply, renewal, use restrictions and hold policies, lost/damaged materials, summary of charging policies, copying charges, billing method and payment methods.

The main impetus for the XML schema project is to record the level of ILL protocol support. Some of the data elements defined in the 1997 draft related to ILL service levels have not been migrated to the schema, probably because of the complexity of the underlying data model.

PRIDE X.500 schema

Australia's Macquarie University is one of the partners in the PRIDE Project – a project funded by the European Commission Libraries Programme with the aim of developing a broker service to support the identification and delivery of information services over the global information infrastructure. The PRIDE Project has decided to adopt the X.500 standard as its framework for directory services. Out of this has come a requirement to define an X.500 profile for libraries which will incorporate a metadata schema for describing libraries, their collections, services and activities. Macquarie University is taking the lead in the development of this profile.

The PRIDE X.500 schema is being developed to support the following scenarios:

The following object classes have been defined within the PRIDE X.500 schema:

A number of objects have also been defined to enable support for the specification of ILL service levels, including ILL charges and lending policies.

Directory Standards and the Australian ILRS Project

In 1998, the National Library of Australia implemented an Australian Libraries Gateway that provides access to information about Australian libraries, their collections, services and activities. The service has an online update capability which enables libraries to be primary partners in maintenance of the currency of the service. The gateway operates simultaneously as a means of collecting appropriate data and of publishing the directory.

Early this year we implemented a Directory of Oral History Collections that builds on the Australian Libraries Gateway Project. We have now begun work on a project to migrate the Australian Interlibrary Resource Sharing Directory, which records interlending policies, conditions and charges, to the same framework.

The national bibliographic utility, Kinetica, has its own user registration module, as does Kinetica Document Delivery. The Council of Australian University Libraries is currently implementing local interlibrary lending systems in Australian universities as an outcome of the LIDDAS project. Each of these has its own directory of suppliers. We would like to enable these services to exploit data that had been created and maintained once only by libraries.

We are working closely with Macquarie University to ensure that these business needs are addressed in a way that will facilitate development of the national information infrastructure and that is aligned with the work of the PRIDE Project. To achieve these objectives we have designed our work program to operate on two levels, one theoretical and the other practical. Outcomes of the ILRS Project will include:

The new ILRS service will extend the data model already developed for the Australian Libraries Gateway. Data elements to describe ILL organisational units and services will be tested against user needs and aligned with the PRIDE schema. ILL service providers will have access to online update facilities. There will be no need to rekey parent organisational details already recorded in the Australian libraries database. Care will be taken to ensure that maintenance of National Union Catalogue symbols is an integral part of the service.

As part of the ILRS Project, an X.500 version of the ILRS database will be built. The X.500 database will have LDIF data interfaces that will enable the sharing of directory data with local resource sharing systems (LIDDAS) and Kinetica Document Delivery.

Parallel with these developments TC 46/SC4/WG7 will be busily rewriting ISO 2146 based on the new ISO 8459 framework. The new version of the standard will be protocol and syntax dependent. It will be based on a concordance with existing standards and work in progress. The focus will be on documenting the data elements needing to be supported by an application to deliver the PRIDE user scenarios.

During March and April, we plan to disseminate a draft set of functional specifications for the new ILRS system widely via the Web. I will also take an early draft of the ISO 2146 work and a report on progress to the next ISO TC46 meeting in Berlin in May. Without committing to a firm delivery date, we hope to have the new ILRS service ready for launch in the third quarter of the year, with operational interfaces to the LIDDAS and Kinetica Document Delivery via the X.500 testbed database.

Standards development can be a slow process but without standards communities may develop systems which do not interoperate when the time comes to share resources in a network environment. Successful projects such as JEDDS have demonstrated the value of addressing immediate business needs in the context of an overarching service model that identifies gaps in existing data standards and network protocols and works to address them. The ILRS Project, with its three parallel streams, attempts to emulate this approach.

As an immediate outcome, improvements to the national information infrastructure should be achieved through mechanisms enabling the maintenance and exchange of directory information supporting the interlibrary loan and document delivery process. From the outset, these mechanisms will support access to directory information through the X.500 protocol. Data elements needed to support these mechanisms will be documented and fed into the standards development process through ISO TC46 SC4 WG 7. Because the standards work was based on practical experience, the new ILRS service should be positioned for enhancement as standards are finalised.

References

BUBL Information Service. BUBL Link 5:15: Selected Internet Resources Covering all Academic Resources. Online. Available: http://bubl.ac.uk/link/. 11 February 2000.

Crossnet Systems Ltd. Analysis of the Z39.50 Profile for Access to Digital Collections and the Z39.50 Explain Service for UKOLN. Crossnet Job No. F17. 29 October 1997. Online (PDF). Available: http://www.ukoln.ac.uk/metadata/cld/study/crossnet/zpadc.pdf . 11 February 2000.

Dublin Core Metadata Initiative. DCT1 - Dublin Core Types 1 (Revised 1999-08-05). Edited by Rebecca Guenther. Online. Available: http://mirror.nla.gov.au/dc/groups/type.htm. 11 February 2000.

Fretwell-Downing Informatics. Z39.50 for Archival Applications: Scoping and Draft Specification Document. Specialised Research Collections in the Humanities. 1996. Online. Available: http://www.kcl.ac.uk/projects/srch/reports/z3950.htm. 11 February 2000.

Heaney, Michael. An Analytical Model of Collections and their Catalogues. Third issue revised. Oxford. 14 January 2000. Online (PDF). RSLP Collection Description. 18 January 2000. Available: http://www.ukoln.ac.uk/metadata/rslp/model/amcc-v31.pdf . 11 February 2000.

IETF Calendaring and Scheduling (calsch) Working Group. Internet Calendaring and Scheduling Core Object Specification (iCalendar) (RFC 2445). Online. Available: http://www.imc.org/rfc2445 . 11 February 2000.

International Council on Archives. ISAAR(CPF): International Standard Archival Authority Record for Corporate Bodies, Persons and Families. Final ICA approved version. Ottawa, 1996. Online. Available: http://data1.archives.ca/ica/cds/isaar_e.html. 11 February 2000.

International Council on Archives. ISAD(G): General International Standard Archival Description. Ottawa, 1994. Online. Available: http://data1.archives.ca/ica/cds/isad(g)e.html. 11 February 2000.

ISO/TC 46/SC 9. N 269: Resolutions of / Résolutions de l'ISO/TC 46 /SC 9. Paris, France. 1999-05-18/19. Online. Available: http://www.nlc-bnc.ca/iso/tc46sc9/docs/sc9n269.htm. 11 February 2000.

Kay, David & Ibbotson, Ian. National Networking Demonstrator Project . Technical Consultants Report. Version 3.0 (final). Specialised Research Collections in the Humanities. November 1998. Online. Available: http://www.kcl.ac.uk/projects/srch/reports/fdi.htm. 11 february 2000.

Leech, Helen. CIRCE: Better Communities through Better Information.. Library and Information Commission research report 1. Online. Microsoft Word. Available: http://www.lic.gov.uk/publications/executivesummaries/rr001.html. 11 February 2000. 11 February 2000.

McCarthy, Gavan. Heritage and the Internet - Encoding Context Objects: Using Knowledge to Reduce Risks. Australian Society of Archivists 1999 Conference. Online. Available: http://www.archivists.org.au/events/conf99/mccarthy.html . 11 February 2000.

North American Interlibrary Loan and Document Delivery (NAILDD) Project. Interlibrary Loan Directory. Draft. December, 1997. Online. Available: http://www.arl.org/access/naildd/ipig/res/ipig9712-gen-inst-info.shtml. 11 February 2000.

Powell, Andy et al. Collection Level Description: A review of existing practice. An eLib supporting study. UKOLN. August 1999. Online. Available: http://www.ukoln.ac.uk/metadata/cld/study/. 11 February 2000.

RSLP Collection Description. UKOLN. Online. Available: http://www.ukoln.ac.uk/metadata/rslp/. 11 February 2000.

Scottish Confederation of University and Research Libraries (SCURL). CAIRNS: Co-operative Academic Information Retrieval Network for Scotland. Online. Available: http://bubl.ac.uk/org/scurl/docs/cairns1.htm. 11 February 2000.

Scottish Confederation of University and Research Libraries (SCURL). Research Collections Online (RCO). Online. Available: http://bubl.ac.uk/org/scurl/. 11 February 2000.

Society of American Archivists. Standards for Archival Description: A Handbook. 1994. Online. Available: http://www.archivists.org/catalog/stds99/index.html. 11 February 2000.

Standards Australia. AS 4590-1999 Interchange of client information. PDF. Available: http://www.standards.com.au/Catalogue/Script/Details.asp?DocN=stds000023947. 11 February 2000.

Sweetmore, Keith. National Networking Demonstrator Project for Archives. A project funded through the NFF Monitoring Programme under the auspices of the JISC. Report of the Archival Consultant. Version 2.0 (final). Specialised Research Collections in the Humanities. November 1998. Online. Available: http://www.kcl.ac.uk/projects/srch/reports/ksfinhtmldoc.html. 11 February 2000.

Top

About This Site | Copyright | Privacy | Accessibility | Site Map | Site Search | Content A-Z | Contact Us