BL RIC logo 3.0 Retrospective Conversion of Library Catalogues in Institutions of Higher education in the UK

(Note. Discussion of the factors relating to retrospective conversion which are covered by this section - e.g. benefits, costs, definitions, overlap - are of general relevance to all the library sectors dealt with by the BLRIC study and therefore this discussion is not repeated in Section 4.)

3.1 Summary of the FIGIT study

The study of the justification for a national programme of retrospective conversion of library catalogues in UK institutions of higher education was commissioned by the Follett Implementation Group on IT (FIGIT) and funded by the Higher Education Funding Councils (HEFCs) through their Joint Information Systems Committee (JISC). The study resulted from a recommendation of the Libraries Review Group chaired by Professor Sir Brian Follett and took place between October 1994 and April 1995. Led by Philip Bryant, it was in two parts. Part A gathered quantitative data and was undertaken by Russell Sweeney (Library Consultant) assisted by Steven Prowse (then Technical Support Officer, UKOLN ).

Part B examined the justification for a national programme through the use of a Focused Consultation Group led by David Streatfield of Information Management Associates. Philip Bryant was assisted with various aspects of the project, in particular those relating to costs and retroconversion methods, by Ann Chapman (Research Officer, UKOLN ). The study was advised by a Monitoring Group chaired by Bernard Naylor (Librarian, University of Southampton).

3.1.1 Main quantitative results

  1. 10,500,000 records (estimated to represent 2,000,000 individual titles) had been converted in UK HE libraries. The number of catalogue records remaining to be converted in the libraries responding to the survey totalled some 28,000,000 - a figure consistent with those given by Hoare (1986) and Law (1990). It is estimated that this figure represents 6,000,000 individual titles.
  2. Of the 28,000,000 records 4,000,000 are for items to be found in 'special collections'; the majority of these being in the Humanities. From information collected much was deduced about the subject and geographic spread of this material.
  3. The costs of converting an existing manual catalogue record to machine readable form fall within a range of 1 to 5, with the mean within the range 1.50 to 2.

3.1.2 Conclusions

  1. The 'justification' for retrospective catalogue conversion is accepted both nationally and internationally.
  2. Users as well as librarians are now concerned about those records which have not been converted - not least for accessing material within the local library.
  3. If a national programme of retrospective conversion for the HE sector were established, some librarians would be prepared to reallocate funding when participating in the programme.
  4. It is clear that there would be a need for 'machinery' to coordinate a national programme and to prevent duplication of effort.
  5. Equitable 'criteria' would have to be agreed regarding access to items in collections.
  6. There is recognition of the importance of organising efficient and fair access to the total national resource of bibliographic records.
  7. There is concern that all converted records should meet bibliographic standards of a level which will ensure effective consultation and exchange.
  8. The 'non-HE sector' is recognised as being of vital importance to the HE sector and investigation of the need for, and problems of, retrospective conversion in this area is required.
  9. To date libraries have met most of the costs of retrospective conversion from current expenditure. There is less likelihood of this being possible in the future. In addition, if the recent Non-Formula Funding allocations are ignored, there has been a significant reduction in external funding.
  10. The British Library must be a major player in any national programme of retrospective catalogue conversion.

3.1.3 Recommendations

  1. A national programme should be agreed, with due regard to the role of the non-HE sector, especially the British Library, and consistent with Section 3.1.4 of this report.
  2. 25,000,000 is required over a five-year period to enable the complete retrospective conversion of records for 'bibliographic' stock. (This figure does not take into account the impact of the Non-Formula Funding allocations already made. See Section 3.2.3).
  3. As a general rule, 'matching' money should be forthcoming from institutions in receipt of 'special' funding.
  4. Machinery should be established for the management of the programme i.e. to coordinate effort, set priorities and target funds.
  5. In relation to funding the following should be considered:
    • 'Preferential pricing agreements' to be sought from record suppliers
    • 'Criteria' to be produced by the library community which could inform decisions regarding possible funding allocations from other sources available nationally.
  6. Clear guidelines should be agreed and made available on:
    • 'Standards' for retrospectively converted records
    • 'Procedures' for retrospective conversion programmes
  7. All records created as a result of a national programme should remain in the public domain and 'principle' 4 of the Recommendation R(89)11 of the Council of Europe Working Party on Retrospective Cataloguing should apply: 'on the basis of reciprocity, converted catalogue records should be able to circulate unrestrictedly (N.B. the word 'freely' was not used) within and between library networks, without legal or contractual constraints on their use by other members of those networks.'
  8. The development of a national database is of prime importance in implementing a national programme of retrospective conversion and discussions which are currently under way between the JISC, CURL and the British Library are welcome in this context.
  9. Further studies should be carried out in the following areas:
    • The scale of the problems relating to non-print library material e.g. manuscripts, photographs, slides.
    • Retrospective conversion in the non-HE sector.

3.1.4 A national programme

Within the higher education community there is ample empirical evidence that, if retrospective conversion is left to the decisions of individual institutions, its achievement will be haphazard and long-drawn out. Since the achievement of 'universal' retrospective conversion will benefit the whole of the centrally funded HE community, it seems essential to promote a co-ordinated national effort. Furthermore the HE community is both the major provider and the major user of the scholarly library resources of the UK . It is therefore a responsibility for the HE sector to make its own major contribution towards unlocking the nation's total resource of scholarly library material.

Since the Sweeney study defined the scale and scope of the task, it is possible to envisage a systematic attack on it, amounting to a 'national programme'. The national programme should have the following features:

  1. An organising and monitoring group should be established to be responsible for allocating funds and monitoring the progress of individual institutions. This group would be responsible for:
    1. vetting applications and for ensuring that the pattern of allocating funds for retroconversion can be accommodated within the budgetary constraints;
    2. taking decisions to ensure that the greatest benefits from the programme are derived at the earliest possible date, and also for promoting the optimum availability of good quality records, for the benefit of all libraries engaged in the programme;
    3. giving recipient libraries reasonably long-term security of funding, to help with the hiring of appropriate staff, and inserting periodic checks to ensure that the funds disbursed are producing the promised benefits;
    4. ensuring that the arrangements for access to and re-use of records created are compatible with the concept of a national programme;
    5. rendering an annual report to the Funding Councils of the monies disbursed and the progress made towards the goal of 'universal' retroconversion.
  2. Institutions should be invited to bid for funds by submitting applications which take due note of the main relevant points covered by this report.
  3. Applications should state the nature and scale of the task, the preferred method of approach, and the timetable it is intended to follow. Specific reference ought to be made to any particular priorities to be followed in the programme and to any local constraints or other factors, helping to shape the particular project.
  4. The availability to other parties of the records produced, both for consultation and for re-use, should be stated. The amount of money requested and the amount of local resource to be contributed should also be stated.

3.2 Background to the study

Retrospective catalogue conversion is one area where it can be said that once money has been invested in the process a permanent benefit is assured. The question is - benefit to whom? Of the 28,000,000 records which remain to be converted in the libraries covered by this study, nearly a half are accounted for by nine of the largest, older university libraries. (ANNEXE 1).

The libraries of the new universities, in the main, already have the catalogue records for their stocks in machine-readable form. These institutions may well have a set of priorities where a programme for the retrospective conversion of other libraries' catalogues is placed well down the list; however, the major thrust of the Libraries Review was to assist new universities to avoid the need to build new research collections.

3.2.1 The wider research community

There are many references in the Follett report to resource sharing and to making access to collections available to researchers 'across the system as a whole'. In addition, Paragraph 27 of the report states: 'a more strategic approach to providing library facilities in support of research in all subjects needs to be developed involving both higher education institutions and other providers of research oriented library and information services'. Within the latter context the present study could not review retrospective conversion issues as a whole, but had to be restricted to the higher education sector. Obviously funding constraints prohibited FIGIT from funding research outside the HE sector. It is, however, recognised that academia and scholarship make heavy demands on the whole range of library provision - national, learned society, public, heritage institution and other.The report of the Group on a National/Regional Strategy for Library Provision for Researchers (1995) chaired by

Professor Michael Anderson is also very relevant here.

'The Group was agreed that such a strategy should involve the active participation of the national copyright libraries, university research libraries, the libraries and resource centres of the Research Councils, the larger public libraries and, preferably, some libraries funded by learned and professional societies'

Although FIGIT was unable to finance a study to include these types of library, the importance of the non-HE sector has been noted and discussions took place with the BLRIC with a view to the Centre funding a further study at the end of 1995.

3.2.2 Humanities research collections

The Follett report (Paragraph 25) made a firm recommendation) regarding the need to preserve and maintain research collections and to provide improved access to them, which resulted in the Non-Formula Funding of Specialised Research Collections in the Humanities. This recommendation was made with the intention that such funding 'should form part of the funding councils' response to the Government's decision not to create a research council for the humanities' A condition to be placed on institutions in receipt of Non-Formula Funding awards was that they would be required to provide free access to all bona fide researchers from within the UK. This was a welcome initiative but it was not coordinated with this study and the results given in Russell Sweeney's report were unable to allow for the impact of grants made by the Panel on Non-formula Funding chaired by Professor Martin Harris. It was of some concern to the study's Project Monitoring Group that HEFCE Circular 5/95 (1995) gave a somewhat restricted definition of 'bona fide researcher' i.e: ... employed in HEIs funded by the Funding Councils and DENI.

3.2.3 Higher education Non-Formula Funding of Special Collections in the Humanities

(Note. The following update on the allocation of money by the Harris Panel has been provided by Ian Mowat, University Librarian, Edinburgh University)

About 50,000,000 was distributed by the Harris Panel to institutions in higher education following the recommendations of the Follett Review Committee to improve access to special collections. Both non-recurring and recurring projects were supported, the latter for up to four years. Three hundred and twenty seven projects in sixty six institutions of higher education throughout the United Kingdom have received support.

Non-recurring funding was given for projects covering conservation, cataloguing and preservation. In addition, recurring funding supported projects designed to ensure increased access through publicity, the development of the collections or support for user-related services. The estimated breakdown is as follows:

Cataloguing ConservationPreservation
111 60 20
Cataloguing ConservationPreservation
140 6719846

The aim of the cataloguing projects primarily is to produce electronically available records for books and manuscripts. An analysis is just about to commence, comparing the projected numbers of records which it was intended to prepare for current output. It is unlikely that these figures will be available within the next few months but the final output, on completion of the projects, ought to number several million records (including many brief records for manuscript material)

OCLC probably has been the largest external supplier of records and, in general, the normal library standards have been applied for book cataloguing. While the great majority of records will be available via existing library OPACS on the net some, for local reasons, will be restricted to PC access at the individual site. A project to investigate the best way of making available all NFF-created records is just about to be funded. It seems unlikely, at this stage, that the records will be added directly to the CURL OPAC (COPAC), but it is possible that a research 'clump' containing both COPAC and NFF records may be created.

The NFF projects are being monitored by a committee serviced by Jacqueline Fitzgerald and Rachel Bruce, Programme Co-ordinators of the Joint Information Systems Committee (JISC) of the Higher Education Funding Councils and chaired by Ian Mowat of Edinburgh University Library. The Committee receives regular reports from projects, has a programme of visits of inspection and provides support in a number of ways. The programme as a whole, though late in starting, is broadly running to schedule and, while some of the initial estimates of work to be done proved rather optimistic, the final outcome will be in sight of the original targets.

The Monitoring Committee is also charged with reporting back to the Funding Councils through the JISC and will be making recommendations on possible next steps for this initiative. Discussions are taking place to coordinate consideration of the work of NFF, the Anderson Report and the retrospective conversion studies.

A guide to the projects Accessing our Humanities Collections, was published in 1997 and is available on request from

Library Office,
Kings College London,
London WC2R 2LS.

Further information is obtainable from and enquiries should be addressed to Jacqueline Fitzgerald at the above office. Tel: 0171-873-2599, Fax: 0171-873-5080 E-mail:

3.2.4 Chris Hunt's paper to the Libraries Review

The impetus for the FIGIT study was provided by an unpublished paper commissioned by the Follett Sub-committee on Information Technology from Chris Hunt (1993), Director and University Librarian, John Rylands Library, University of Manchester. It was this paper which included among its recommendations the suggestions for Non-formula Funding and the need for a feasibility study of a national retrospective conversion programme.

The following is a selection from the points made by Chris Hunt:-

The suggestion of a feasibility study was taken up and provided the impetus for the recommendation for the FIGIT study given in the Follett report (Paragraph 364)

3.2.5 Reasons for retrospective conversion - local benefits

The Follett report (Paragraph 302) recognised that: 'Much unique material, particularly early books and books printed outside the UK, is often still accessible only through the manual catalogue systems of libraries. Both Hunt's paper and the Follett report quite properly focused on the shared benefits of retrospective conversion, but it should be stressed that it is probable that the great majority of retrospective conversion programmes have not primarily been motivated by concern for other members of the wider library community, but rather by the perceived benefit for the local library and its users. A frustration often facing users of academic and research libraries is that they cannot search in just one place - the online catalogue - to gain knowledge of what is in their own library's stock, but have to consult another form of catalogue - sometimes several. These other catalogues are not always as accessible as they should be, sometimes being housed out of the immediate public gaze.

In Russell Sweeney's study (ANNEXE 1), in response to the question 'What form of non-machine readable catalogue are still in use?' the results were:


An extreme example is a major university research library where users can find themselves consulting up to eight files in all five manual forms plus three online files. It is not surprising therefore that, given the rapid development of online catalogues, users and librarians alike are wanting to have data on their own library's stock available in just one place and are becoming increasingly concerned about catalogue records which have not been converted to machine readable form. This is a point which emerged, not only in the discussions of the Focused Consultation Group (ANNEXE 2), but also from the experience of some members of the Project Monitoring Group. Improved awareness of the contents of the home library makes economic sense from at least three points of view:

3.2.6 Reasons for retrospective conversion - resource sharing

As far as resource sharing is concerned there are three main reasons for undertaking retrospective conversion:

  1. To benefit scholarship by providing locations, not previously available, for unique or unusual items.
  2. To share the load of supplying items which are too frequently demanded from just one or two institutions because the locations of additional copies have not been known.
  3. Regional and national cooperation is facilitated because retrospective conversion enables sensible management decisions to be made in relation to acquisition, preservation and withdrawal of stock.

With respect to these reasons it has become clear from all the discussions which have taken place is that the British Library (BL) must be a major player in any significant national initiative taken with regard to retrospective catalogue conversion, both to ensure that effort is not duplicated and also to enable the BL to plan and develop services with more informed knowledge of resources available nationally. In addition it is important that the new Library and Information Commission and the Library and Information Cooperation Council (LINC) be involved in developments.

3.2.7 Should all records be converted?

There are many little used items of apparently low importance in libraries' stocks and concern is expressed in some quarters that these items do not justify money spent on the conversion of their records. This concern and the value judgements it reflects (one person's 'essential research' item is another person's 'ephemeral' item) lie at the very heart of any debate about the nature and purpose of libraries. It is not the role of this report to expand on this debate; however, it is perhaps worth making two points:

  1. At some time in the past these items have been considered sufficiently important to have had the manual catalogue records created for them in the first place.
  2. The experience of members of the Focused Consultation Group, of the Project Monitoring Group and of those undertaking the FIGIT study was that 'use' of stock increases in relation to its 'visibility'.

If there is a national OPAC containing catalogue records created and funded as part of a national programme of retrospective conversion then libraries must ensure that copies of the titles represented by those records are retained in stock.

3.3 The study

3.3.1 General approach

The objectives of the project were as stated in the Follett report (Paragraph 303) and were to establish the following:

  1. How much retrospective conversion of research collections has been undertaken.
  2. What remains to be done.
  3. The benefits that would accrue to users from national investment in this activity.
  4. Evidence of demand for access to such material which is not being met at present.
  5. Identification of:
    1. costs
    2. methods of assisting the process
    3. sources of funding that might be drawn upon for such activities.

It was agreed that the project should be in two parts. The first part would gather quantitative data relating specifically to Objectives (i) and (ii) and would also address some aspects of (v). The second part, addressing Objectives (iii) and (iv) would examine the justification for national funding for retrospective conversion and, as value judgements would be involved, a qualitative research approach was required.

Part A: Establishment of quantitative data relating to retrospective conversion in the UK HE sector was undertaken as a questionnaire survey by Russell Sweeney, a Library consultant, and Steven Prowse, then on the staff of UKOLN. (ANNEXE 1)

Part B: A qualitative study examining the justification for a national programme for retrospective conversion was undertaken as a Focused Consultation Group led by David Streatfield and Graham Robertson of Information Management Associates. (ANNEXE 2)

The objectives as identified by the Libraries Review picked up a number of the issues given in Chris Hunt's (1993) paper; nevertheless, although the aim and objectives of the project as proposed in the Follett report were succinctly stated, a number of relevant matters were either not mentioned, or else were not made explicit. A few examples taken from those given in the project proposal were:

3.3.2 Material covered by the quantitative survey

As with the previous studies by Peter Hoare (1986) and Derek Law (1988) the focus of this study was on bibliographic material because the scale of the retrospective conversion task in relation to these items is definable in broad terms and can be seen as finite. The problem of covering material in all formats, however, is a huge one. Although the Title II-C Program in the US has funded work on material in all formats, it is not clear how much of the money spent on this other material has been actually dedicated to improving bibliographic access compared to preservation and collection development. From the Washington discussions in 1994 it was apparent that the aspect of retrospective conversion now of most concern to the US academic/research community is non-print library material. Positive interest was expressed in discussing mutual problems in this area with the UK.

3.3.3 Defining what has to be counted

The library community has long had problems with the production of consistent and comparable statistics. The production of a comprehensive list of consistent and internationally accepted terms would be of immense value. Statistics are only useful when there is agreement about what is being counted and costed. Nowhere has the problem been greater than in the areas of cataloguing and the quantifying of stock. Three terms especially cause difficulties; title - record - item. For the purpose of Part A of the FIGIT study the following was agreed for inclusion in the 'Notes for completion' of the questionnaire:

Title. A named bibliographic entity which is catalogued as a unit, forming the basis for a single catalogue record. A title will be represented in a collection by one or more title occurrences, called items.

(Example. Shirer's The rise and fall of the Third Reich is one title represented by one record although a library holding it may stock ten copies (items). If 100 libraries stock this title they will each create/acquire a record resulting in 100 records for the one title and they may stock 200 copies (items) between them. Considerable economies would be achievable if 99 of the libraries could derive their records from one 'source' record.)

The first sentence of the above definition was used again in the 'Notes for completion' of the BLRIC study's questionnaire. However, there were still problems with librarians varying understanding of what should be counted or estimated and these problems are referred to again in Section 4.5.6.

3.3.4 Overlap between stocks

It is recognised that if items are unique or unusual the conversion of the records for those titles benefits scholarship, but questions are often asked about the degree of overlap between stocks. Hoare (1986) wrote 'the question of overlap in holdings is clearly important, since a low level of common stock reduces the economic attraction of cooperation. More overlap between stocks means that more libraries have the opportunity to save money on the acquisition/creation of catalogue records. If the catalogues of the larger libraries are converted first then the smaller libraries can benefit. In addition, the greater number of known locations for the titles held should ensure that demand on individual libraries and wear and tear on the stocks involved are more equitably shared.

There has been very little study of overlap in the UK since that undertaken by the University of Lancaster's Library Research Unit in 1971 on behalf of the National ADP Study (1972). The Project Monitoring Group asked that a major UK database be approached to see if the system could generate information to show:

The current study was pleased to receive cooperation from LASER who kindly agreed to analyse their files. The analysis showed that, excluding the BLDSC files held by LASER, 18,000,000 locations represented 2,400,000 titles of which 900,000 were 'unique'. In the unlikely circumstance of there being a direct correlation between the LASER figures and those of the pool of records awaiting conversion in UK HE library catalogues, the 28,000,000 records would represent 3,700,000 titles of which 1,400,000 would be unique. However, the fact that the LASER database contains locations for many public and other libraries as well as for academic libraries means that the degree of overlap is certainly higher. In addition, 4,000,000 of the records to be converted relate to 'special collections' where the degree of overlap will be significantly less. It seems reasonable to assume therefore that the 28,000,000 records probably represents 6,000,000 individual titles, with a greater proportion of these being unique than in the case of the LASER database.

3.3.5 Costs

It was recognised from the beginning that the establishment of precise costs by use of a self-completed questionnaire was not possible. It was for this reason that the original project proposal suggested that the costs element of Part A of the study should be treated as a separate exercise. However, a decision was taken not to proceed with this idea at the time as it was anticipated that the British Library might be prepared to support a broader based cost study of bibliographic record creation and management in due course. It was hoped that libraries responding to the questionnaire survey would be able to provide sufficiently accurate figures from their management systems, but when Russell Sweeney 'piloted' the questionnaire and followed it up with a number of face to face interviews he found that the questions which posed the greatest problems were those relating to costs. This situation was also clearly repeated in the BLRIC study.

In the FIGIT study libraries were often unable to separate the costs of any retrospective conversion from those for their current cataloguing. Analysis of questionnaires in the full survey demonstrated this fact further; nevertheless, the libraries' excellent response rate in Part A of the study allows the assumption to be made that the mean unit cost figures arrived at can be accepted as generally satisfactory for the purpose of estimating funding needs.

In addition to the survey data, information was sought from the principal bibliographic record suppliers and is as follows:-


Every retrospective conversion service has its own method of charging. Charges can vary considerably depending on the complexity of the original catalogue. There can be discounts for contribution of new records to a database, or for members of cooperatives. Libraries may decide to use one retrospective conversion service to undertake the whole project, or they may choose to split the records to be converted into sections and use different services for different sections.

The figures that follow summarise the current charges of a number of major services.

Tag and key
This uses a printed form of the catalogue (cards, photocopies of cards, shelf lists, etc.) which is marked up for keying in. The keying in is usually done by a data preparation bureau or the data preparation section of the retrospective conversion service.
Costs: 1.10 - 1.50.
Add on costs: photocopy of original catalogues, staff to mark up entries. The retroconversion service will do both at an additional cost.
Searching databases online
An operator checks each record against the database to see if a match occurs.
  1. By retrospective conversion service
    Costs: 1.00 - 1.20 per record found.
    Add on costs: supplying data for search terms (ISBN or other); the cost of sending the original catalogue, or providing a copy; adding local data (if not added by a retrospective conversion service).
  2. By libraries themselves
    Costs: 16p - 35p per record found, up to 74p per search (whether record is found or not).
    Add on costs: staff adding local data and editing entries if required; also purchase of extra equipment; service charges per terminal used; special software packages.
Searching database offline
A file of keys created from the original catalogue is run against the database of the retrospective conversion service.
Costs: 16p - 35p per record.
Add on costs: supplying data for search terms (ISBN or other); adding local data (if not added by retrospective conversion service).
Full retrospective conversion service
This service will use a combination of methods to convert the records required by the library.
Costs: 1.00 - 2.00 per record. Could be up to 3.00 a record.
Upgrading existing machine readable records
This is often the conversion of short title records, e.g. circulation records which are too brief, and/or records in an unsuitable format for current use.
  1. A specially written program is used to convert records from one form to another.
    Costs: up to 1.00 per record.
  2. The old record file is used as the basis of an online or offline search.
    Costs: as for online or offline searching.
Image scanning
This procedure can only be used on certain forms of catalogue, where the images are clear and uniform in layout. The cards are image scanned, machine readable records are produced, and previously agreed index points are tagged by the system.
No UK work done so far. Costs: 40p - 50p per record (estimated from overseas sources).
Further 'add on' costs
Equipment:Specific hardware for retrospective conversion
Dedicated terminals for retrospective conversion
Software:Packages for online/offline searching by libraries
Service charges
Photocopying of cards or shelf lists for keying in or for search term lists.
Other factors to be considered
  1. Best hit rate for the library in question
  2. Required type of record
  3. Most appropriate method of producing new records for unmatched titles.

An invitation was also extended over LIS-SCONUL to librarians who had relevant costs data to provide it. Half a dozen libraries responded most helpfully.

The conclusion was that there was a broad range of costs due to a variety of factors, such as the characteristics of the material for which the records were being converted; the availability of machine readable records from external sources; the quality of the original catalogue records. It was calculated that the cost of converting any particular record lies within the range of 1 to 5 with the 'mean' being within the range 1.50 to 2. This information was given in a letter sent as an 'interim report' to the JISC Secretariat in November 1994.

If, as is recommended, all 28,000,000 records are converted and libraries receiving grants contribute 50% of the cost (except in special cases) 25,000,000 of special funding would be required. This should constitute a five-year programme, the aim being to complete the retroconversion of all 'bibliographic' records by the end of the millennium. This sum, which does not take into account allocations already made under Non-Formula Funding, represents less than half of the money spent on UK university recurrent expenditure on books and periodicals during 1993-94 (Universities Statistical Record, 1995).

3.3.6 Tackling retrospective conversion

The first of the priority issues identified by the Focused Consultation Group was resources; this heading covered both 'overt' and 'covert' costs. In relation to 'covert' costs matters were raised such as space for extra staff employed and the capacity of the library's IT infrastructure to cope. In addition, other factors influence how many records a library can buy with its money. A major consideration is whether a library is going to carry out retrospective conversion with the 'book in hand' and undertake so much editing that the process becomes one of retrospective cataloguing rather than retrospective conversion. The 'technical report' published with the Council of Europe's Recommendation (R(89)11) stated that:

'In order to minimise project costs of retroconversion the catalogue records should not be edited ... before or during the conversion ... Editing should be less expensive when carried out after standard retroconversion because it can be expected that a growing supply of machine readable records from other sources can be used for improving one's own records'

Russell Sweeney, in eleven follow-up interviews undertaken after his survey found that, in all cases except one, new records are created from existing records with recourse to the item itself only occurring in cases of difficulty.

Many librarians are daunted by the prospect of having to cope with the extra work posed by undertaking a program of retrospective conversion and integrating that programme with the current flow of their library's routine activities. There are a number of sets of guidelines to help librarians define the precise aims of their projects; to identify the problems; to consider the range of options open to them. The major bibliographic record supply agencies will give good advice and in at least one instance a clear set of guidelines is published.

Two sets of guidelines independent of any commercial interest have also been produced. The first is LITC Report No. 4, originally published in 1992 and updated and reissued by the Library and Information Technology Centre (1994) at South Bank University. This set is in two parts. Part 1 gives a review of the options and Part 2 gives details of specific record suppliers and services. The second set, prepared by Anton Bossers and Derek Law on behalf of the LIBER Library Automation Group (1990), was published as an annex to the Technical Report issued with the Council of Europe's Recommendation (R(89)11).

These latter guidelines make the important point that a realistic timetable should be set and that there should be regular monitoring of progress during the course of any programme to ensure 'timely adjustment if necessary'.