BIBLINK - LB 4034 - D1.1 Metadata Formats

Metadata Formats
Work Package 1 of Telematics for Libraries project BIBLINK (LB 4034)

6. Metadata Requirements of National Libraries

As part of this work package a questionnaire was circulated to five national libraries: Bibliotheque Nationale de France (BNF), Koninklijke Bibliotheek of the Netherlands (KB), the Norwegian National Library (NB), Biblioteca Nacional (BN) in Spain and the British Library (BL). The full responses are attached in Annex 2, and a consolidated report is included in this section. The questionnaire was designed to gather some initial information about the cataloguing practices of the various national libraries regarding electronic publications, and their future requirements.

The responses appear to support the view that the metadata format(s) adopted for electronic resources are consistent with those adopted for conventional print material. There are obvious benefits to existing national bibliographic agencies if the full range of document types can be handled in a system with a common structure. However if separate services were to be provided solely applicable to electronic resources this would no longer be the case.

The responses indicate that libraries wish to create full records describing objects at a detailed level in terms of bibliographic content, terms and conditions, access, subject content, location etc. They intend to create records in the various flavours of MARC currently used for print publications. They intend to apply detailed cataloguing rules to the content of the records. The records created in this way are envisaged as forming the basis of a range of services involving selection, retrieval and access, record supply and preservation.

As part of the consensus building process we need to establish libraries' requirements as regards the level of detail supplied by publishers. Are libraries seeking detailed records or is the intention to add value by enhancing simple records? Do libraries want detailed records embedded in electronic publications for use at the cataloguing stage or would they prefer simple records to be supplied in advance of publication? Possibly libraries will only ever want simple records for those items which are not legally deposited or for particular categories of publication.

The timing within the publication process of supply of metadata from publisher to library has a crucial impact on the format of metadata. For off-line publications from 'traditional' electronic publishers, national libraries may be wanting to use information received from the publisher in advance of gaining access to the object. In this case libraries are looking for the equivalent of a CIP system. Either there could be a one-off supply of 'minimal set' metadata supplied pre-publication, or additional detailed metadata could also be supplied as it became available. For many newer web based publications speedy creation of metadata is of the essence, the aim is to minimise the delay between appearance of the electronic resource and provision of metadata.

In this context Lenart identified significant factors in the supply and demand of records as:

delay in supply
quality of records
subject indexing
legal limitations imposed on re-use of records
charges payable to rights holders.

(Models for the provision of national bibliographic services in Europe : final report / by Michele Lenart ; translated and edited by the British Library. - Luxembourg : European Commission, 1996).

These factors are worth considering in the context of electronic publications and can be interpreted as:

at what stage of the publishing process are records required by national libraries?
at these identified stages what quality of record is required? (The required level of detail and granularity may differ)
what added value is required at the various stages of the process? (For example inclusion of subject indexing, adherence to cataloguing rules, adherence to authority files may vary)
what are the implications for ownership of records? (If publishers supply detailed records they may wish to retain ownership of these records).

As regards this last point, the business model will differ significantly from country to country. Lenart identified re-use and copyright issues as significant in the following countries: Belgium, Greece, Ireland, UK, Portugal and Italy. However the cost implications of other factors (level of detail, timing of record supply, added value) would be significant in all business models.

Another issue for consideration is the disparity between publishers and libraries in their use of rules for formulation of content. Typically publishers do not follow cataloguing rules for content even in detailed SGML headers. Will complex information supplied by publishers be of benefit to libraries if it needs to be re-formulated to adhere to cataloguing rules? As part of the refinement of scoping project partners need to consider how far they will guide publishers as regards rules for formulation of the content of the information they supply.

It is worth noting that three libraries (BNF, KB and NB) indicate they are using the Dublin Core elements as a way of exploring means to catalogue electronic resources.

The advantages of embedded metadata as opposed to 'free-standing' metadata are alluded to, one significant aspect of this is authentication. We will leave this issue to the authentication work package.

6.1 Responses to the Questionnaire

All of the five libraries have had some experience in cataloging off-line publications. These are deposited (either legally or voluntarily) in the Netherlands, France, Spain and Norway. In the United Kingdom consideration is being given to extending deposit laws to non-print items. KB and NB have started to create records for on-line publications.

6.1.1 Cataloguing format and cataloguing rules used currently

Q1. Please indicate the cataloguing system you use and the cataloging rules you apply.

Responses indicate that each country uses its own cataloguing rules, all based on ISBD, and its own flavour of MARC, which will vary more significantly.

KB PICA PICA rules are ISBD based. Guidelines for electronic documents are under development.

BL UKMARC AACR2.

BN IBERMARC Spanish Cataloging Rules (based on ISBD).

BNF INTERMARC AFNOR (National French Standardisation body) standards for cataloguing rules. Plus ISBD(CF) based rules for electronic documents.

NB BIBSYS-MARC based on USMARC Norwegian Cataloging Rules (based on AACR2) plus a Norwegian supplement for off-line machine readable files and an OCLC guide for on-line.

Q2. What data elements do you record for an electronic publication in addition to those you record for a traditional publication?

Q3. Which MARC (or equivalent) fields do you use to hold the data element

Q4. What data would you like to include in the record that you cannot find, or have difficulty finding from the publication?

Q7. What metadata elements do you consider will be required for electronic publications? Examples: description of content (including mention of sub-units, contents pages); relation to other documents

This table lists all the data elements mentioned in response to questions 3, 4 and 7 as well as question 2. It therefore includes:

elements already recorded in cataloging electronic publications
elements that should be recorded but can be difficult to find
elements it is felt will need to be recorded in the future.

Where an element was mentioned by only one library this has been noted in the last column.

Data Element Comment

personal author Person primarily responsible for the intellectual content.

other contributors Statements of responsibility for multiple contributions.

corporate author

definitive title Variants of the title can appear on boxes, accompanying information and internal sources - these often conflict.

unique identifier e.g. ISBN.

place of publication

publisher Agency responsible for producing the publication.

host Agency making the publication available - e.g. SURFnet.

date, year Date of publication.

exact date of issue To distinguish updates in electronic resources.

price For off-line publications.

language

edition Particularly significant for on-line publications.

update information "

version information "

general material designation e.g. [computer file].

specific material designation e.g. tape, diskette etc.

type of computer file e.g. data, program etc.

file characteristics e.g. size, number of records contained.

additional information e.g. sound, image, text, multimedia etc.

relationship to printed versions Existence of printed versions e.g. a digitised novel.

subject keywords Need for controlled vocabulary. Cataloguers may use publishers suggestions as a guide.

legal deposit number

classification
BN

description of content More detail than keywords alone would assist selection and acquisition by users of BNB. Summary, tables of contents. BL (BNB)

Additional data for on-line publications

availability Free of charge or by account.

terms and conditions

login name and password

file format HTML, pds, ps, wp51, etc.

URL

file name
elements defined by KB

path
"

number of files, bytes
"

compression format
"

type of connection
"

port number/protocol
"

gopher type
"

name of computer - host
"

IP address computer - host
"

location - host
"

mail address - host
"

mail address - person
"

URL and other details as above for images linked with publication To present an image which is linked with the document as part of the bibliographic description. "

service provider

links A listing of links to other documents. NB

unique title
NB

duration of availability at given URL
BNF

Additional data for serials

frequency How often it will appear.

regularity Information relating to the production of new articles. Will they only appear at with the next issue, or as and when they are ready for publication? BL

data relating to articles Information relating articles to the journals they appear in.

abstracts

contents page data See also, 'description of content' in the main table. BL

System requirements

Statement indicating if specified system requirements are preferred or required.

Agreement on standard place in publication and/or in accompanying documentation for system requirements.

off-line Processor, memory, operating system, application software, monitor, cards, peripherals etc.

installation and de-installation information

on-line Internet browser, viewer, telnet client, FTP client, WWW client etc.

Other related data

Amount of time taken to install This information will be important for access to the documents by the end-user - it needs to be taken into consideration for the access service. BNF

local address For off-line or remote material made available on a local network.

Q3. See 2 above

Q4. What data would you like to include in the record that you cannot find, or have difficulty finding from the publication?

The individual elements mentioned in response to this question have been included in the tables above. The following notes include some additional remarks.

BNF, BL, KB Difficulties finding the source of the description. In electronic publications the information is scattered throughout the document and can take some time to find.

BL, BN, BNF Statements of responsibility: a CD-ROM can have many different functions : script, developer, infographist, designer, title manager, music, etc.

KB, BN, BL, BNF, NB Technical data and system requirements: for installation and de-installation for cataloging; for access now; for access in the future. Distinguish between required and preferred hardware and software.
It would be easier if all this was presented in a standard format and location in a publication.

KB, BNF, NB Terms and conditions/access information.

BL Publisher name and location.

BL Correct title - variants on boxes, accompanying information and internal sources often conflict.

BL, BN Version/edition information. (Is a serial a first issue?)

BNF Precise date of publication.

NB The 'size' of an on-line document. What (and how many) files/records does it consist of? A listing of links to other documents.

BN ISBN and legal deposit numbers are not shown on disc labels or elsewhere.

Q5. How have you resolved or attempted to resolve the difficulty?

All the libraries had some experience of cataloging off-line publications although at some this was still at the developmental stage, and for some publications, relied on external sources. Off-line publications are treated very much as printed books or serials with extra data added relating to system requirements.

KB establishing test beds.

BL developing rules to apply to problems as cataloguing rules do not let you assume anything.

BN contacting publishers for further information.

NB contacting other libraries to see how they have resolved the problem.

asking advice of IT personnel.

Q6. What library services will use metadata for electronic publications?

KB OPAC (local retrieval system).
Union Catalogue (NCC) (national retrieval system + ILL).
Delivery of electronic resources e.g. published by universities (WebDOC).
List of electronic publications as a by-product of the national bibliography.
(possibly in the future) alerting services.

BL Facilitating access and record supply
Selection
Acquisition
Preservation and archiving

BN Most library services will use metadata.
A problem to be considered with deposited electronic publications is copyright and a possible unfair use.

BNF National Bibliographic Agency both for legal deposit and National Bibliography.
Specific bibliographic by-products for electronic documents
Services in charge of access to the collections.

NB All library catalogues
The metadata will probably have to be refined to meet the requirements of the national bibliography.

Q7. What elements will be needed in future? See 2 above

Q8. Will the records for electronic publications need to be integrated in your existing systems?

KB Yes, the records of electronic publications produced in the PICA format are in the same database with the other title records. It must be possible to have both kinds of records in the OPAC's of the library and e.g. in files of the National Bibliography. A total integration is required.

BL As far as practicable and to the same basic bibliographic and format standards. The British Library is aiming for uniform access to its collections.

BN Yes. The National Library wants to have a single database.

BNF Off-line electronic publications are already integrated in our system. For on-line, it will be the same. We will have a unique multimedia integrated system for the OPAC and national bibliography. This requirement is very important because we will adapt the existing system to new electronic documents but we will not completely change our system or our format in the next five years.

NB Yes.

Q9. Will the metadata need to be manipulated by particular protocols e.g. Z39.50, EDI, etc.?

KB PICA exploits since the end of 1995 a Z39.50 entry to the central database GGC. In view of this it is important to investigate the pros and cons of manipulating metadata by the Z39.50 protocol in comparison with the use of the PICA protocol.
Nowadays we buy tapes with bibliographic records which are integrated in the common database maintained by PICA. As output, tapes are sent to our CD-ROM production company in the United States. Possibilities to use FTP for transfer of the records of the national bibliography for the CD-ROM are investigated.

BL Probably MARC, Z39.50, HTML.

BN Yes. At least, Z39.50, ISO 10162/10163, EDI, and any other protocols established by the DG.XIII.

BNF Yes, the catalogue will support the Z39-50 protocol. We plan to give access possibilities to other libraries to the BNF national bibliography records via FTP.

NB Yes, certainly by Z39.50. And by other relevant protocols (even if we do not use them e.g. the EDI protocol).

Q10. Further comments about experiences in cataloging electronic publications.

KB
KB
Off-line electronic publications are catalogued and published in the national bibliography.
The cataloguing of CD-ROM and diskettes does not present specific problems. Except for the difficulty in some cases to judge whether a hybrid publication (book + diskette) is mainly a printed or mainly an electronic publication.
Concerning on-line resources, especially home pages: At this moment three cataloguers are making bibliographic descriptions for home pages. One for the scientific collection, one for the union catalogue and one for the national bibliography. Further, since a common cataloguing system is used they can also extract records from the common database and they can see how cataloguers in other organisations, especially one university library catalogue e.g. home pages. The problems with home pages are: author unknown, publisher unknown, date of publication unknown, change or even disappearance of the URL.

BL In terms of physical carriers, a cataloguer may be presented with any number of formats - CD-ROMs, floppy disks of two different sizes, magnetic tapes, cartridges, all of which may be specific to different platforms (PC, Mac, Acorn, etc.). It will be necessary to be in possession of the equipment and software needed to access the publication.
It takes considerably more time to create a record for an electronic publication due to: 1) installation and de-installation 2) procedures to give internal access may entail reading more of the document than in paper publications.
One of the systems issues raised by the project is the difficulty in having simultaneous access to the WLN cataloguing system, and the item in hand. The ability to "hot-key" between the item's title screen and the cataloguing form screen is essential.

BN We have little experience of this type of publication but we are very interested in all problems with this material.

BNF BNF has a solid experience for cataloguing off-line electronic publications. The on-line topic is very important to us and we will benefit from BIBLINK results before starting a national project.
The French national bibliography is based on the legislation for legal deposit which includes off-line electronic publications. The problem is to include progressively on-line publications but then we come back to the limits. What do we include: electronic journals, home pages (on which criteria etc).
The evolution towards cataloguing "articles" is also an important issue as we do not include articles in the national bibliography.

Q11. Wish list - anything else?

KB For the National Bibliography and other metadata products and services of the library, the consistent and user-friendly presentation of records on screen or on paper is very important. In the international library community a proposal for an ISBD(CF), the presentation of metadata of computer files, has been commented upon. In December 1996 the final version is expected. In the meantime we have defined ourselves an 'ISBD'-like presentation for 'On-line Resources' and PICA has implemented this in its system.
Linking between different versions of electronic publications.
Links between journal and articles of the journal (whole/part relationship) should also be realised in the system.

BL Every item (off-line) should have a de-installation file.
There should be a standard place and format for including installation instructions in the user handbooks accompanying the item.

NB The quality of the metadata elements should be ensured, especially by controlled vocabulary, e.g. name of persons and institutions, subject keywords.

Next Table of Contents

KB	PICA	PICA rules are ISBD based. Guidelines for electronic documents are under development.
BL	UKMARC	AACR2.
BN	IBERMARC	Spanish Cataloging Rules (based on ISBD).
BNF	INTERMARC	AFNOR (National French Standardisation body) standards for cataloguing rules. Plus ISBD(CF) based rules for electronic documents.
NB	BIBSYS-MARC based on USMARC	Norwegian Cataloging Rules (based on AACR2) plus a Norwegian supplement for off-line machine readable files and an OCLC guide for on-line.

Data Element	Comment
personal author	Person primarily responsible for the intellectual content.
other contributors	Statements of responsibility for multiple contributions.
corporate author
definitive title	Variants of the title can appear on boxes, accompanying information and internal sources - these often conflict.
unique identifier	e.g. ISBN.
place of publication
publisher	Agency responsible for producing the publication.
host	Agency making the publication available - e.g. SURFnet.
date, year	Date of publication.
exact date of issue	To distinguish updates in electronic resources.
price	For off-line publications.
language
edition	Particularly significant for on-line publications.
update information	"
version information	"
general material designation	e.g. [computer file].
specific material designation	e.g. tape, diskette etc.
type of computer file	e.g. data, program etc.
file characteristics	e.g. size, number of records contained.
additional information	e.g. sound, image, text, multimedia etc.
relationship to printed versions	Existence of printed versions e.g. a digitised novel.
subject keywords	Need for controlled vocabulary. Cataloguers may use publishers suggestions as a guide.
legal deposit number
classification			BN
description of content	More detail than keywords alone would assist selection and acquisition by users of BNB. Summary, tables of contents.		BL (BNB)


Additional data for on-line publications
availability	Free of charge or by account.
terms and conditions
login name and password
file format	HTML, pds, ps, wp51, etc.
URL
file name		elements defined by KB
path			"
number of files, bytes			"
compression format			"
type of connection			"
port number/protocol			"
gopher type			"
name of computer - host			"
IP address computer - host			"
location - host			"
mail address - host			"
mail address - person			"
URL and other details as above for images linked with publication	To present an image which is linked with the document as part of the bibliographic description.		"
service provider
links	A listing of links to other documents.		NB
unique title			NB
duration of availability at given URL			BNF

Additional data for serials
frequency	How often it will appear.
regularity	Information relating to the production of new articles. Will they only appear at with the next issue, or as and when they are ready for publication?	BL
data relating to articles	Information relating articles to the journals they appear in.
abstracts
contents page data	See also, 'description of content' in the main table.	BL

System requirements
Statement indicating if specified system requirements are preferred or required.
Agreement on standard place in publication and/or in accompanying documentation for system requirements.
off-line	Processor, memory, operating system, application software, monitor, cards, peripherals etc.
installation and de-installation information
on-line	Internet browser, viewer, telnet client, FTP client, WWW client etc.

Other related data
Amount of time taken to install	This information will be important for access to the documents by the end-user - it needs to be taken into consideration for the access service.	BNF
local address	For off-line or remote material made available on a local network.

BNF, BL, KB	Difficulties finding the source of the description. In electronic publications the information is scattered throughout the document and can take some time to find.
BL, BN, BNF	Statements of responsibility: a CD-ROM can have many different functions : script, developer, infographist, designer, title manager, music, etc.
KB, BN, BL, BNF, NB	Technical data and system requirements: for installation and de-installation for cataloging; for access now; for access in the future. Distinguish between required and preferred hardware and software. It would be easier if all this was presented in a standard format and location in a publication.
KB, BNF, NB	Terms and conditions/access information.
BL	Publisher name and location.
BL	Correct title - variants on boxes, accompanying information and internal sources often conflict.
BL, BN	Version/edition information. (Is a serial a first issue?)
BNF	Precise date of publication.
NB	The 'size' of an on-line document. What (and how many) files/records does it consist of? A listing of links to other documents.
BN	ISBN and legal deposit numbers are not shown on disc labels or elsewhere.

KB	establishing test beds.
BL	developing rules to apply to problems as cataloguing rules do not let you assume anything.
BN	contacting publishers for further information.
NB	contacting other libraries to see how they have resolved the problem.
	asking advice of IT personnel.

KB	OPAC (local retrieval system). Union Catalogue (NCC) (national retrieval system + ILL). Delivery of electronic resources e.g. published by universities (WebDOC). List of electronic publications as a by-product of the national bibliography. (possibly in the future) alerting services.
BL	Facilitating access and record supply Selection Acquisition Preservation and archiving
BN	Most library services will use metadata. A problem to be considered with deposited electronic publications is copyright and a possible unfair use.
BNF	National Bibliographic Agency both for legal deposit and National Bibliography. Specific bibliographic by-products for electronic documents Services in charge of access to the collections.
NB	All library catalogues The metadata will probably have to be refined to meet the requirements of the national bibliography.

KB	Yes, the records of electronic publications produced in the PICA format are in the same database with the other title records. It must be possible to have both kinds of records in the OPAC's of the library and e.g. in files of the National Bibliography. A total integration is required.
BL	As far as practicable and to the same basic bibliographic and format standards. The British Library is aiming for uniform access to its collections.
BN	Yes. The National Library wants to have a single database.
BNF	Off-line electronic publications are already integrated in our system. For on-line, it will be the same. We will have a unique multimedia integrated system for the OPAC and national bibliography. This requirement is very important because we will adapt the existing system to new electronic documents but we will not completely change our system or our format in the next five years.
NB	Yes.

KB	PICA exploits since the end of 1995 a Z39.50 entry to the central database GGC. In view of this it is important to investigate the pros and cons of manipulating metadata by the Z39.50 protocol in comparison with the use of the PICA protocol. Nowadays we buy tapes with bibliographic records which are integrated in the common database maintained by PICA. As output, tapes are sent to our CD-ROM production company in the United States. Possibilities to use FTP for transfer of the records of the national bibliography for the CD-ROM are investigated.
BL	Probably MARC, Z39.50, HTML.
BN	Yes. At least, Z39.50, ISO 10162/10163, EDI, and any other protocols established by the DG.XIII.
BNF	Yes, the catalogue will support the Z39-50 protocol. We plan to give access possibilities to other libraries to the BNF national bibliography records via FTP.
NB	Yes, certainly by Z39.50. And by other relevant protocols (even if we do not use them e.g. the EDI protocol).

KB KB	Off-line electronic publications are catalogued and published in the national bibliography. The cataloguing of CD-ROM and diskettes does not present specific problems. Except for the difficulty in some cases to judge whether a hybrid publication (book + diskette) is mainly a printed or mainly an electronic publication. Concerning on-line resources, especially home pages: At this moment three cataloguers are making bibliographic descriptions for home pages. One for the scientific collection, one for the union catalogue and one for the national bibliography. Further, since a common cataloguing system is used they can also extract records from the common database and they can see how cataloguers in other organisations, especially one university library catalogue e.g. home pages. The problems with home pages are: author unknown, publisher unknown, date of publication unknown, change or even disappearance of the URL.
BL	In terms of physical carriers, a cataloguer may be presented with any number of formats - CD-ROMs, floppy disks of two different sizes, magnetic tapes, cartridges, all of which may be specific to different platforms (PC, Mac, Acorn, etc.). It will be necessary to be in possession of the equipment and software needed to access the publication. It takes considerably more time to create a record for an electronic publication due to: 1) installation and de-installation 2) procedures to give internal access may entail reading more of the document than in paper publications. One of the systems issues raised by the project is the difficulty in having simultaneous access to the WLN cataloguing system, and the item in hand. The ability to "hot-key" between the item's title screen and the cataloguing form screen is essential.
BN	We have little experience of this type of publication but we are very interested in all problems with this material.
BNF	BNF has a solid experience for cataloguing off-line electronic publications. The on-line topic is very important to us and we will benefit from BIBLINK results before starting a national project. The French national bibliography is based on the legislation for legal deposit which includes off-line electronic publications. The problem is to include progressively on-line publications but then we come back to the limits. What do we include: electronic journals, home pages (on which criteria etc). The evolution towards cataloguing "articles" is also an important issue as we do not include articles in the national bibliography.