Loafmann, 'Retrospective Conversion for Special Formats: AV and Microforms', LITA Newsletter v14n04 URL = http://hegel.lib.ncsu.edu/stacks/serials/lita/lita-v14n04-loafmann-retrospective [v14n4.recon litanews] ---------------- Retrospective Conversion for Special Formats: AV and Microforms Kathryn Loafmann AV Retrospective Conversion: Planning for the Conversion Project Cynthia M. Whitacre (OCLC) began this program, sponsored by the LITA/ALCTS Retrospective Conversion Interest Group. Cost is a consid- eration in planning any retrospective conversion project, but even more so for AV (audiovisual materials). AV retrospective conversions are more costly because of the complexity of the material, which re- sults in increased costs for training and searching. The scope of pro- ject will also influence the final cost. Will the project be limited to sound recordings, films and videos; or do you intend to include games, realia, and kits? The first point to consider is the source file. To convert means that some sort of printed record exists that will be converted to ma- chine-readable form. If the printed form is a catalog card, were the records created under AACR2? If not, do you want the records converted to AACR2? Are these standard catalog records or are these nonstandard records that happen to be on catalog cards? The printed form may be in some format provided by the Media Library: a book catalog, list or printout from a database. The entries may not look like catalog re- cords, but do contain the minimum amount of information needed to iden- tify or create a record: title, publisher, date and physical descrip- tion. Examine the source file for local practice information. If the pub- lisher/distributor information was entered in code in the source file, do you have the code list? You will need to specify exactly where the information is stored in the source-file records. You must be consistent. For example, what field contains the local call number: the 099 field or a 9xx field? You will need to make decisions about format; for example, the pres- ence or absence of brackets in the 245 subfield $h for the general ma- terial designation (GMD). Do you want the vendor to supply the brack- ets? The way to "do it once and do it right the first time" is to record and document the decisions at each step in the planning process. The next point to consider is your definition for a matching re- cord. How flexible are you willing to be in identifying a match? What fields in the MARC record do you want to examine and compare to the en- tries in your source file? How will you deal with multiple matches? If your conversion will be done by a vendor, ask the vendor if they have a document on the procedure for identifying a matching record. Such a document can be used as a starting point for developing your own match- ing procedure. For example OCLC uses these fields to determine a match: 008 (Type, Bib lvl, and Type of material); title proper (245 $a); edition, if present (250); publisher (260 $b, 261 in older re- cords); date of publication (260 $c, 261 $d in older records); and physical description (300, may extend into 5xx fields). There are critical points for determining a match for each type of AV material. For visual materials they are number of physical units, length (important for films), format (important for videos), color or black-and-white, sound or silent, negative or positive, size (impor- tant for films, as in 36mm or 16mm), and accompanying material. For sound recordings, they are music publisher's number, form (12" disk or cassette or CD disk) and mono or stereo. For music, they are the per- forming group and the composer. The third point to consider is the expected hit rate. This will be affected by the amount of variation you are willing to accept in the record content. For videos, timing is an area of considerable variance; allow ten minutes variation for the length of the video. Some records were en- tered with an estimate of the length of the video, while other records contain the exact length based on viewing the entire video with a stop watch. Another area of variance for videos is dates. The date may be for the original movie or it may be the date that the video was dis- tributed. If all else matches, allow a three year variance for dates. Choosing a right record for a kit is difficult. Local practice in the definition of what constitutes a kit has resulted in a wide vari- ation in the GMDs input for kits. The GMD on the source file entry may say kit, but the actual item might be a game or a filmstrip. The best policy is to be very liberal in matching for kits, and say that any AV record will be an acceptable match. Use the planning phase of the project to identify where in the source file you have problems, and determine what matching strategy you are going to use in order to maximize the hit rate. Choosing the right record requires looking at a number of records and making a choice. There will be multiple records for forms of the same item, and many will be within the same format. For this reason, batch processing is not recommended for AV materials. The matching for AV material should be done by people who can exercise judgment in iden- tifying matching records. The final point to consider is what editing you will want done. What do you want the records to look like? Do you want an 007 field en- tered? Do you want the correct GMD? With or without brackets? Do you want the subject headings from the source file entered? The summary notes? Do you want authority control to be done by a vendor? Extensive editing is time consuming and will increase the cost of the project. What local information do you need to have entered? Bar codes, call numbers, location codes, price, accession number, copy and volume in- formation are all possibilities. You will need to define the formats for this information and identify in which fields you want the informa- tion entered. At the end of the project, you will be left with the titles that did not have a match. Do you want to create a new record for no-match cases? Do you want to have these entered by a vendor from the source file? Should the vendor enter the records under the latest cataloging rules, or the rules in effect at the time the source record was cre- ated? Would you rather handle these as part of your regular internal workflow and recatalog from the piece? Retrospective Conversion for Special Formats: Major Microform Sets Birong Ho (Wayne State University Libraries) spoke on Wayne State's ex- periences. Wayne State University Libraries had a team working exclu- sively on retrospective conversion since 1984 using OCLC MICROCON, but found the in-house retrospective conversion going very slowly. In 1988 they began the Kellogg Project (so named because Kellogg of Battle Creek was funding it) which was intended to address all the as-yet-un- converted materials, including government documents and major micro- form sets. The retrospective conversion of the major microform sets held by Wayne State took five years to complete. The first step was to identify the major microform sets owned by the institution, then to check this information against the OCLC list of major microform sets. Wayne State selected thirteen sets from the list of OCLC major microform sets to be loaded into their online cata- log. It was discovered, after the tapes were loaded, that one of the sets loaded was not the same as the set held even though they had iden- tical titles. Their experience would suggest first creating a test file of each tape before loading the tape into the local system in or- der to be sure it is the right microform set. Also discovered after the tapes were loaded was the presence of a number of duplicate records in the online catalog. Technical Services had been cataloging selected microform titles at the request of the Reference Department. Preservation programs had also been creating bib- liographic records for these titles. There were also some duplicate bibliographic records on the tapes (for the set Early English Books, there were 3,181 duplicate records) that were detected and not loaded. On the other hand, records were loaded for material that was miss- ing or filed incorrectly. If the material couldn't be located, it was necessary to suppress the record. Loader programs were used by the Systems Office to load the records into the database. Call numbers and universal messages were added to the records as part of batch jobs done at time of load. As a result of this batch job approach, there were some confusing messages in the notes field about where the material was filed. Differences in how the call numbers were constructed among the mi- croform sets posed a problem. For the set Early English Books, the reel number was used to file the set and the call number was derived by combining information from the 830 and 510 fields. For other sets the actual enumeration on the piece was the call number. In other cases, the call number was derived solely from the series title field (830 or 440). There were some problems with some of the records--for example, ar- ticles in the title subfields of 6xx and 7xx fields, and series titles entered in varying formats. There were also discrepancies between ti- tle proper as it appeared in the bibliographic record and the title as it appeared on the piece. For example, in the bibliographic record the title appeared as History of the Chaldaic philosophy but on the piece the title appeared as History of the Chaldaick philosophy. Authority control was contracted to Blackwell North American (BNA) for preprocessing. There were a number of situations that authority control preprocessing didn't handle. Subdivisions for corporate bodies were not recognized. Names lacking dates that were entered as subject tracings were not caught. Typographical errors in headings were not caught. Variant forms not appearing as cross-references on the rele- vant authority record were not caught. Personal names were a particu- lar problem, since many were old forms of names for which there were no authority records. Upon request, vendors can supply a variety of reports useful in cleanup activities. Authority capture/processing can produce lists for updating. BNA could have supplied automatic alerting service for changed/deleted authority records, but Wayne State didn't use this be- cause of cost considerations. Local systems can also generate selected reports by using a report generator. The load program will also create reports that can be used in authority work: lists of new name and subject headings to database, deleted name and subject headings, and conflict and error detection re- ports. A list of headings generated because the headings could not be indexed can be useful for resolving subfield coding errors. Loader re- ports can indicate records with incorrect locations, too many fields, fields too long, or required fields missing. Index creation also gener- ates error reports. A standard number report, such as one based on OCLC numbers, will spot duplicate records. A call number report can spot call number formatting errors. Another useful standard number re- port is the item index report. The thing to bear in mind is that er- rors can enter the database through a tapeload that would otherwise be blocked at time of inputting. As a result, you will see problems that you never encountered before. In the final stages, however, the identification of records needing cleanup was a manual process. Librarians working in authority control examined categories of headings based on prior experience with problem headings. They also relied on reports from the public services staffs. One final task, called the Shelflist Project, was performed after the cleanup project was done. The shelflist card was compared to the record online, and local information was added to the records. Finally, some insights from the question-and-answer period: *Be prepared for boxes of microform to turn up after project is over-- or even during the project. *Quick and dirty planning means a lot of cleanup work later. *Lack of a standard number on AV material makes retrieval difficult.-- Kathryn Loafmann is at the University of North Texas Library