MARBI Members:
Priscilla Caplan LITA Univ of Chicago James E. Crooks RASD Univ of Calif, Irvine Shelby Harkin ALCTS Univ of North Dakota Sandra Lindberg RASD Lafayette Public Library Jacqueline Riley RASD Univ of Cincinnati Paul Weiss ALCTS Univ of New Mexico Larry Woods LITA Univ of Iowa Robin Wendler LITA Harvard University Jacqueline Riley RASD Univ of Cincinnati MARBI Interns: Elaine Henjum LITA Florida Ct. For Lib Auto Carol Penka RASD University of Illinois Representatives and Liaisons: Joe Altimus RLG Research Libraries Group John Attig OLAC Penn State Univ Randall K. Barry LC LC, Net Dev/MSO Betsy Cowart WLN WLN Bonnie Dede SAC Univ of Michigan John Espley AVIAC VTLS Catherine Gerhart CC:DA Univ of Washington David Goldberg NAL NAL Richard Greene OCLC OCLC Rebecca Guenther LC LC, Net Dev/MSO Michael Johnson MicroLif Follett Maureen Killeen ISM ISM Library Info Services Rhonda Lawrence AALL UCLA Law Library Karen Little MLA Univ of Louisville Sally McCallum LC LC, Net Dev/MSO Susan Moore MAGERT Univ of Northern Iowa Young-Hee Queinnec NLC NLC Marti Scheel NLM National Lib of Medicine Louise Sevold PLA/CIS Cuyahoga County Pub Lib Patricia D. Siska ARLIS Frick Art Ref Lib Rutherford Witthus SAA Univ of Colorado Other attendees: John Albright RTIS Joan Aliprand Research Libraries Group Karen Anspach EOS International Kathy Bales Research Libraries Group Connie Barrencleigh Health Canada Sara Ramser Beck St. Horn's (?) Public Library Julia Blixrud CAPCON Library Network Francis Bondurant New York University Arevig Caprielian New York Society Library Ann Case H.W. Wilson Co. Sherman Clarke New York University Karen Coyle Univ of Calif Division of Lib Auto Carroll Davis Columbia University Michael Fox Minnesota Historical Society William W. Jones New York University Lola Halprin Emory University Ruth Haas Harvard University Diane Hillmann Cornell Law Library Jean Hirons Library of Congress Charles Husbands Harvard University Sherry Kelley Smithsonian Institution Libraries Rhoda Kesselman Princeton University Kris Kiesling RLG / SAA-CAIE Mary Larsgaard Univ of California, Santa Barbara Elizabeth Lilkes Brooklyn Historical Society Clifford Lynch Univ of California Div of Lib Auto Andrew MacFowan The British Library Elizabeth Mangan LC / G&M Ingrid Parent National Library of Canada Cecilia Preston ---- Patricia Rader New York Public Library Becky Ringler Univ of Calif, San Diego Barbara Roberts DC Public Library Frank Sadowski University of Rochester Jacque-Lynne Schulman Natl. Library of Medicine Louise Sevold Cuyahoga County Public Library H. Shoemoh (?) British Library Thom Shudargas CCLA / Tallahassee Ann Sittirn Health Canada Sandia Smalley Harvard University Gary Smith OCLC Stuart Spore New York University Medical Steven Squires Univ of North Carolina at Chapel Hill Dianne Stalker State Univ of New York at Stony Brook Barbara Story LC Vitus Tang Stanford University Susanne Warren Art & Architecture Thesaurus Bob Warwick Rutgers University William Wilson New York University Bob Warwick Rutgers University Jill D. Yaples Binghamton University Libraries MARBI Recorder: Josephine Crawford University of Wisconsin, Madison ***************************************************************** Saturday, July 6, 1996 ***************************************************************** Priscilla Caplan opened the meeting with introductions and an announcement about a planned MARBI Character Subset meeting. --------------------------- DISCUSSION PAPER NO 96 : Defining a Uniform Resource Name Field in the USMARC Bibliographic Format. Rebecca Guenther introduced the first discussion paper on Uniform Resource Names (URNs). She explained that this paper is more "FYI" than for any serious discussions leading to MARBI decisions in the near future. She suggested that it might be useful to post links from LC's web site to the important IETF (Internet Engineering Task Force) documents. Cliff Lynch brought some up-to-date news from the recent Internet Task Force (ITF) meeting, since the discussion paper was written before the meeting was held. He said that the original URN group was disbanded and the charter of a new group was well received. Cliff expects a URN syntax paper by December; he thinks this paper will provide enough flexibility to incorporate existing schemes. It is likely that an existing system, e.g. ISGN, could be embedded into a URN scheme. Cliff also expects a long and painful standards process to hammer out the resolution issues. One or two resolution schemes will be worked on separately by the new group. Apparently, the handles scheme has not been well received enough to date to be considered, and PURLs (from OCLC) are considered to be a form of URL. The primary scheme under consideration is a hybrid of Ron Daniels' work and the Domain Name Server scheme. Priscilla thought it would be helpful to have a review before proceeding. A URN is a consistent name regardless of the location of the electronic resource or network object. A URN is only useful when it can be resolved, meaning: -- your Internet request returns one or more locations. -- your Internet request returns metadata about your object. -- your Internet request returns your object itself. MARBI needs to be thinking about: -- what should library users get back? All of the above, or does it depend upon the situation? -- what is the relationship to cataloging records? Paul Weiss stated a preference for using the 0xx field for the URN rather than the 856 field. Karen Coyle agreed; she advised a repeatable field at the bib record level. Cliff felt that there is no clear answer: there are multiple and complex relationships involved. How much should libraries try to track? Paul pointed out that the 020 has similar implementation problems. Priscilla remarked that different naming agencies will have their own rules. The question of single or multiple records will be wide open. Should $8 be used for linking purposes? Rebecca discussed various naming authority possibilities within the library community --LC, an ALA body, NISO, or other. Joe Altmus wondered how serials would be handled. If there is no connection between the number of URNs and cataloging records, will a URN change occur when the publisher changes but the title does not? Rebecca believes that publishers will not follow a single scheme themselves. Paul said that we have been dealing with similar problems but on a smaller scale. Cliff encouraged the library community to hash out the intellectual content issues related to URN. He said that the IETF has no expertise or interest in this area. The library and publishing communities should work together. John Attig had two points to make. First, some things we already do can become URNs, although the URN syntax may require additional information. Second, when there are multiple URNs for an object, catalogers must choose which to use. Rebecca said this is a time for librarians to be proactive rather than reactive. The bibliographic community can establish its own naming authorities and rules on such situations. No one was quite sure who/what this should be though. It was agreed that MARBI should promote discussions on URNs. ---------------------------- DISCUSSION PAPER NO 97 : Coding Digital Items in Leader/06 (Type of record) in the USMARC Bibliographic Format There have been various related discussions in the past, the most prominent involving a recent Maps proposal. This DP seems to emphasize cataloger flexibility dependent upon various factors, such as form and user needs. The paper supplies a proposed revised definition for Leader/06 Computer File "m" code. John Attig suggested discussing the big picture first -- he believes that the issue is wider than digitized materials. Too many materials are a valid member of two or more categories and it does not serve the users for catalogers to select one only. He sees systems relying upon the leader/06 for processing when this is not always helpful. His example was a limit in OPACs. Robin Wendler believes it depends upon the situation. Sometimes a primary limit is helpful, other times multiple limits for the same object is better. Rich Greene reported that the Leader/06 is used very heavily by the OCLC de-duplication programs. It would be a major effort to change these programs. He agreed with Robin that the situation varies from the user perspective. Karen Coyle reported that the Melvyl programs also do much duplicate detection; it would be much harder since the leader/06 was an optional field. Priscilla wondered if the Leader/06 will be needed once multiple versions are always on the same record. Does it boil down to a content versus carrier issue? Randy Barry said that the analogy to microform breaks down with executable computer files. Priscilla agreed; since there is a one-to-one relationship between the Leader/06 and the 006 fields, there is a fatal flaw if the Leader/06 "m" definition deals with executables only. Priscilla suggested that the cataloging community work on guidelines, while MARBI keeps the USMARC definition more open ended. Paul Weiss said that the OCLC documentation does give some direction now. For instance, it tells catalogers what to do when cataloging something like a map puzzle (two cataloging aspects). He said that the concept is good, but that detailed guidelines are difficult to formulate. Catherine Gerhart agreed to take the issue to the ALCTS CCS CC:DA Committee. Priscilla asked if there was a consensus to open up the Leader/06 computer file definition. Betsy Mangan asked if MARBI could recommend that content be coded in the 008, while the carrier is coded in the 006. Would this be possible? Randy thought that the 007 must also be considered. Sally reported that originally MARBI tried to keep physical description codes in the 007 and the bibliographic description codes in the 008. She suggested that if an 007 is used, perhaps the 006 would be unnecessary. But, unfortunately, there is the legacy of the past to consider. The group discussed in some detail what types of material ought to be coded under a revised definition of Leader/06 Computer Files code. Paul believes it is appropriate for executable software. Robin said that Harvard wants to use this code for statistical data or social science datasets (the latter usually don't have print equivalents). Another suggested expanding the use to raw computer data. What about kits? What about source code? What about files where both software and data are combined? It was suggested that issues for digital material be dealt with separately from other combinations. Participants requested that the discussion paper come back as a proposal with two options: one a narrow option to define code m as just executables and a broad option to define m to include executables, data sets, and raw non-numeric data. It was also recommended to consider revising the definition of kit and look at code p for mixed materials. It was agreed that the issue needs more analysis and discussion, and everyone was encouraged to post examples on the USMARC list. ---------------------------- PROPOSAL 96-7 : "Changes to FTP File Label Specifications for Electronic Files of USMARC Records" Sally McCallum introduced this proposal. It was discussed as a discussion paper at the last meeting, when RLG in particular had some particularly helpful comments. COBRA is a grouping of seven national libraries in Europe; if these changes are adopted by MARBI, COBRA would promote the use of this FTP standard in Europe. Priscilla asked first if there were any general comments. Karen Coyle stated that she was surprised that the proposal covered SGML since she does not see this as a communications format; was the theory that libraries would use for USMARC file exchange only? Sally responded that a Document Type Definition (DTD) is being worked on now to make it possible to transfer a MARC SGML record. Someone asked about file names. Comments on their structure were part of the earlier Discussion Paper, but is no longer needed. Priscilla then began to run through each of the specific proposals: 1. Change label file character set. There was a little discussion about the OR character. No objections registered against this proposal. 2. Change end-of-field (EOF) character symbol. Marti Scheel asked if systems wouldn't use the recommended character in the data itself. This does happen according to OCLC. Rich Greene asked why the proposed standard calls for having two marks (# followed by carriage return or carriage return/line feed), since this seems redundant. Sally said that LC prefers a line feed only. Gary Smith thought that most FTP programs will be able to adjust to the standard. Randy Barry asked if a line can be too long, and a line feed would be needed on its own. Priscilla wondered if there could be a fatal flaw in using the carriage return/line feed alone. A system might end up with just a carriage return, since the line feed is optional; what would then happen if the data itself includes a carriage return? This might cause a problem. Is the # mark really important to COBRA? Unknown. Robin summarized the point of the discussion by saying that a Carriage Return alone was too simple and an end-of-field character was too much. There did not appear to be a consensus at this point in the meeting. Either CR alone or CR + line feed could be used. 3. Add optional Country Identifier field. No discussion. No objections. 4. Make the Format Field mandatory. There was a little discussion, but no objections. 5. Add optional Format Qualifier Field. Gary Smith asked why not use an existing field. Sally responded that the advantage of adding a new optional field is that there will be minimal change to existing programs. Gary suggested adding to the end of the FOR field. Sally said that Gary suggests a more elegant solution, but that it would be hard to do in a fixed field. In general, the working principle with Proposal 96-7 was to modify as little as possible of the existing fields to make implementation easier for all concerned. 6. Add optional Character Set fields. Paul felt that the example was confusing. Little other discussion occurred. No objections. 7. Add optional Character Variation fields. No discussion. No objections. 8. Add optional Final Destination Identification field. Paul suggested that this be a higher label. Perhaps change position? The order of data elements in the past has been mandatory before optional, and fixed length before variable length. Therefore, Priscilla had assumed the FORmat field would move up once it became mandatory. However, others in the room had not assumed this. After discussion, there was general agreement that the order is not important because programs will look for the presence of tags. Paul Weiss moved that MARBI accept the proposal with a single change: drop the # mark as part of the end-of-field marker. Define the end-of-field character to be carriage return or carriage return followed by a line feed. There were six votes in favor and no votes against. ------------------ PROPOSAL 96-8 : CAN/MARC Changes for Marc Format Alignment Sally explained that she would like to introduce Proposal 96-8 today, so that Ingrid Parent (Directeur General, Acquisitions and Bibliographic Service, National Library of Canada) could say a few words on behalf of the National Library of Canada (NLC). The detailed discussion would occur on Sunday and Monday with Young-Hee Queinnec serving as the NLC representative. Sally reviewed the process to date: two discussion papers and impact statements have proceeded the proposal. The statements in particular were very helpful in formulating the proposal. Even though the process of alignment is a difficult one, there are benefits for all in terms of record sharing and less complex systems. Unfortunately, Canada has a particular problem using CONSER records due to CanMARC/USMARC differences and alignment would improve this situation. Sally noted that archival data elements have been excluded from 96-8 and will be dealt with at a later time. She didn't expect a vote on everything but would like for there to be significant progress. Ingrid Parent thanked Sally for her welcome and said she was glad to be attending her first MARBI meeting. She reviewed the history of the MARC harmonization effort, in terms of the British Library initiatives and the general interest in cataloging simplification. She pointed out that CanMARC and USMARC are not very different compared to differences with the other MARC formats. Unfortunately, resolving differences is not an easy task due to the complex systems and large databases involved. Ingrid asked MARBI members to keep in mind that there had already been much negotiation and compromise. What was brought to MARBI in the form of 96-8 does not show the fifty or so changes which Canadian libraries have volunteered to do already, because Canada's user base is smaller. Unlike the U.S. community, the Canadian bibliographic community has to meet bilingual requirements in the creation of cataloging records. English and French are both official languages by law. 9xx linking fields are used for this purpose, but should not be considered "local" fields. They have been part of the Canadian standard all along (with hindsight, another set of fields might have been selected). MARC format harmonization or alignment is an opportunity to make the USMARC format international in scope. There will be much work to accomplish this, but the benefits are even greater. Priscilla asked Ingrid about the many options stated in the proposal. Are all options acceptable to Canada? Not at all. Young-Hee will describe the Canadian preferences in later discussions. John asked how the UKMARC project is going. Sally said that she expects to deliver a discussion paper at the '97 Midwinter meeting followed by a proposal at the Annual Conference. Will any of the decisions here have to be reworked when the UKMARC proposal is under discussion? Sally thinks not. Most areas don't seem to overlap. The one exception might be the 016/010 fields. Robin asked if the British have used the 9xx fields. Sally thought this might be the case. John Attig noted that bilingual support is a big topic, considering both authority and bibliographic formats. Marti suggested we all think "multi-lingual support" instead. Priscilla asked where the archivists are in their work on CanMARC/USMARC harmonization. Rutherford Witthus reported that the Canadians have been in discussion for some time. By the end of the summer, much discussion within the Society of American Archivists should have occurred. There is also a plan to meet with vendors. He expects a discussion paper by the '97 Midwinter Meeting, which will include a definition of archival control. ---------------- BUSINESS MEETING: There were a number of announcements about new LC USMARC products (Update 2 to Bib Format; new Organization Code List edition (formerly Symbols of American Libraries) are out; Community Information Update #1 is ready for printing.) Sally reported that there are many calls for an up-to- date Concise publication. Office staff are in the process of marking up the Concise in SGML, adding examples in the process. The new version will be put up on the USMARC Web site, and also offered (as in the past) as a print publication. Sally discussed a USMARC Web site on documentation status. The printed updates are once a year or when needed, and there is sometimes a need to know what the approved changes are, so it is planned to update the Web status document just after each meeting. Sally reported that the MARC SGML Document Type Definition has been received from the contractor. The purpose is to be able to translate from USMARC to SGML MARC. LC is now contracting for a converter program. The DTD and the converter (when completed) will be available from the USMARC homepage. Actually, there is one Bib DTD (has bib, holdings, and community information) and one Authority DTD. There is a ten-page explanation at the beginning. There are many comments through out. It is a long DTD because generic subfields were not used and it does include obsolete data elements. Sally suggested that the DTDs be considered an "alpha" version, since MARBI might then want to form a working group for a close review. The expectation is that the end result will be a new USMARC standard. There was a question from the audience about progress on LCCNs. Will there be an internal LC workaround due to the century change? The first internal change has been implemented in splitting the Books file at LC. Sally reported that, because LC staff are involved in a major effort to purchase an integrated system, LCCN plans are only at an idea stage at this time. LC is thinking that a four- digit LCCN would be implemented. Sally emphasized that comments about impact are still usefule. There probably will not be a MARBI proposal. Sally raised the interesting question of the future name of the format. USMARC will not work after the harmonization project is completed. Sally and others have considered on IMARC (I can stand for International, Information, Internet-Marc, etc.). What will be the governance for international MARC? This is under discussion as part of the British and Canadian harmonization work. Probably each country will keep their own advisory process in place. There may then be an IMARC Advisory Group where each of the three involved countries (US, UK, and Canada) can veto a proposal. What about the other countries like Australia, Spain, and Switzerland? This hasn't been addressed as of yet. The conversation then moved on to discuss the Classification Format. It has been designated "provisional" for some time now. LC has about 700,000 records to date. There has been only minor tinkering with the original provisional format to date. Is it now time to remove the provisional designation?. What is happening with non-LC classifications? The Dewey people are planning to distribute Dewey tables using the USMARC format. The National Library of Medicine is not planning on doing the same thing with the NLM classification scheme, because it may be abolished all together. Are there any objections to removing the provisional status? Paul Weiss presented some concerns that there are major structural problems with the USMARC classification format. Bonnie Dede suggested looking at international classification needs. It was acknowledged that the mechanism for moving from provisional to full status is not clear. Larry Woods moved to remove the provisional designation from the classification format. Paul Weiss provided a second. There were three votes in favor and one vote against. ***************************************************************** Sunday, July 7, 1996 ***************************************************************** PROPOSAL 96-10: USMARC Character Set Issues and Mapping to Unicode Larry Woods introduced Proposal 96-10, which is the result of several years work by numerous people and an earlier discussion paper. Working principles were communicated earlier and are summarized in the Proposal. Round-trip or bi-directional mapping is really important due to the mixed timing on implementation of the Unicode Standard across systems. Mapping has been completed for the majority of characters, as listed in the paper. Problems and issues have been identified. Larry asked everyone in the room to make a correction on page 3: change F7 from "LEFT HOOK WITH TAIL" to "LEFT HOOK COMMA BELOW." Larry defined "ASCII clone" to be various digits and punctuation marks duplicated in the Latin, Cyrillic, Hebrew, and Arabic USMARC character sets. The Unicode Standard provides only one standard value with which to encode each of these characters. The Standard includes an algorithm for the display of text which accommodates ASCII punctuation marks and digits. If the UNICODE conventions are used, it is not possible to map a string of characters from USMARC to the Unicode Standard and back to USMARC (round trip mapping), and to guarantee that the output will be identical. Larry reviewed three options for dealing with this problem: 1) map to some near equivalent (could lose data in round-trip or bi-directional processing) 2) map to private use space (each system programs) 3) map to Latin equivalents but precede and end with a special non-displaying character (new option not described in Proposal 96-10). Priscilla stated that the primary issue before MARBI members is whether or not the proposed mappings are acceptable. The secondary issue is to see if there is a preference for one of the options for dealing with the ASCII clone round-trip-mapping problem. Joan Aliprand would like to see rigorous evaluation of the three options for handling ASCII clones. If Option #1 is chosen, RLG would have to rewrite current programs. Joan asked for a working principle of "don't modify data unbeknownst to the data owner." Larry gave an example of the bi-directional mapping problem. You have a character string in Hebrew encoded in USMARC and this character string begins with an opening parenthetical mark. The character string is translated to Unicode as the bib record moves to another system. Then it is translated back to USMARC as it moves again. Because the translation is not perfect, the parenthetical mark could curve the wrong way to the user afterwards. Larry would like to see all the options explored in more depth, particularly the third option. Sally encouraged vendors and systems to participate in the analysis, since the RLG contributions have been so helpful. John Espley agreed to ask system vendors for impact statements at the Monday AVIAC meeting. John Attig asked about the work involved in implementing Option #2 (using private use space). Would each site have to program each use of the private space, representing significant work across the bibliographic community? Yes. A few others agreed that Option #2 was not good and would like to know more about Option #3. Priscilla asked if there is a consensus that Option #2 is not a good option. Yes, there was a consensus. One member of the audience pointed out that we cannot assume that users will have a library character set available on their workstations. Another member asked about the property of right-to-left (as in reading Hebrew). How will this be conveyed in bi-directional mapping? Not answered. Robin Wendler asked if only libraries have to deal with these mapping problems. Randy Barry saw the problem as preserving current coding while the computer world is in transition to the Unicode Standard. He expects the problems to be solved eventually as standards fall into place across systems. Paul Weiss emphasized that the problem of round-trip mapping is not an inherent problem with the Unicode Standard. It involves mapping to/from USMARC. He believes the problem is due to a shortcut that was taken when the original ASCII clones were set up in USMARC. Priscilla asked if there was any discussion on the proposed mappings in Appendix 1. There was a question from the audience on adding some badly-needed new characters to the USMARC character set. Paul Weiss stated that it had been agreed to deal with additions later. Joan Aliprand said that RLG is using a superset of USMARC characters and can therefore accept additions. In answer to a question about the next step in the process, Sally said that two things will happen. First, more languages will be mapped. Second, a Unicode MARC record will be set up for a test. Why were Unicode precomposed characters (letters with accents) not used in the mappings or allowed as alternate encodings for the USMARC two-character encodings? Randy said that he has tried to determine the universe of letter-with-diacritic combinations and it was to large-- besides which, Unicode did not have equivalents for all the possible combinations. Joan explained that precompositions will be dealt with in a later phase of the Unicode project. The audience again raised the question of expanding the set of characters in the USMARC format. Several urged that this be done. Joan thinks it should be straightforward to map these, but Randy Barry had a different experience when he worked on it. He said it was harder to map two codes to one, than to map one to one. Sally ended the discussion by thanking the Subcommittee for all its hard work and by saying that she expects the USMARC format to have an expanded character repertoire in the future. Paul Weiss moved that Proposal 96-10 be approved, with the exception of the ASCII clone issue. Shelby Harkin provided a second. There were five votes in favor and one abstention. ------------------ DISCUSSION PAPER #95 : Enhancements to field 007 (Map) for Remote-Sensing Images in the Bibliographic and Holdings Formats Sally McCallum briefly introduced the paper which discusses the need for more 007 codes for remote-sensing images. Susan Moore, representing MAGERT, provided more detail on why codes for the altitude and attitude of a remote-sensing device, amount of cloud cover, characteristics of the sensing platform, and recording technique would be helpful to users. John Attig suggested a separate 007 field for remote- sensing images, rather than additions to the Map 007. Priscilla asked if remote sensing can create data files. Mary Larsgaard responded: yes, the data collected can be place into data files. They are, however, still considered cartographic. Paul Weiss asked why use the 007 for these codes? Cloud cover is not a physical characteristic of the image itself, but indicates something about the content. Apparently, cloud cover information often tells the user about the quality of the image. The amount of cloud cover can impact upon the "usefulness" of the image. Therefore, one might say it indirectly describes condition. Rich Greene reported that OCLC staff like RLG's suggestion to look at a separate 007 field. It is important to know how much of the Maps 007 field would transfer over. Randy had been thinking along the same lines, that a separate field might be better. Susan said that the data elements are not related to the 034 at all. Paul asked if all the needed data elements are now known. He prefers selecting a variable rather than a fixed field, so that additions are easier. Sally agreed that the 007 is supposed to be about the item from a physical point of view, rather than bibliographic (008). She acknowledged that, while these distinctions are generally maintained, it is not 100%. John Attig brought up the Canadian aspect in regards to this discussion paper. CanMARC contains six of the proposed data elements in a local 009 field. As part of format harmonization, Canada would like to move the coding into a non-local field. Priscilla asked if there was a consensus present. There was agreement that adding codes to an existing fixed 007 field is not desirable. There was agreement that the data elements must be accommodated. It was agreed to explore other options, such as a new 007 field or a new variable field. ------------------ PROPOSAL 96-8 : CAN/MARC Changes for Marc Format Alignment The Proposal had been introduced the day before. There are just over twenty different changes, all of which are to be voted on separately. Due to some people's schedules, it was agreed to deal with the map proposals first. PROPOSAL 96-8/007-01 (Globe): Susan Moore reviewed the two codes requested by Canada, "e" for lunar globe and "u" for unknown. She reported that MAGERT supports the request, even though a small number of records will have to be converted. Paul Weiss moved in favor of the Proposal and Larry Woods provided a second. There were six votes in favor of the motion and no votes against it. PROPOSAL 96-8/008-18-21 (Maps): Susan Moore explained the two proposed changes. First, it is desired to change the name of the code "c" so that it is clearer to catalogers. Second, it is desired to add an "m" code for rock drawings. Priscilla asked if there were strong feelings against the name change and new code. No. Paul moved to accept, and Larry seconded. There were seven votes in favor of the motion, and no votes against it. PROPOSAL 96-8/008-22-23 (Maps): Canada requests the addition of three new codes dealing with projection types: az Azimuthal, other type bz Cylindrical, other type cz Conic, other type In preparing the analysis for this Proposal, LC staff looked at the records that had previously been coded "zz" for Other. Of the approximately 180 different projections which have been coded as "zz", only two additional new codes are apparent, Chamberline trimetric projection ("an" suggested) and Lambert conformal projection ("dl" suggested). Young-Hee Queinnec reported that Canada prefers Option 1 (adding five codes and not making any other codes obsolete) because it is less change and less work. Option 2 involves adding the same five codes but also making three codes obsolete. Susan reported that the U.S. map community had no strong feelings in either direction. Priscilla asked if anyone wanted to speak up in favor of Option 2. No one. Paul Weiss moved that Option 1 be approved, and Larry Woods provided a second. Seven members voted in favor of the motion and no one voted against. PROPOSAL 96-8/008-24,25 (Maps): CanMARC has 39 prime meridian values, using both character positions 24 and 25 for them. Six of the 7 USMARC prime meridians are also in the CanMARC list, but with different codes. Canada requests that the USMARC values for positions 24 and 25 be abandoned and the 39 2-character codes in CanMARC be adopted. Susan Moore explained that the prime meridian can be any longitude, so the list of possibilities is almost infinite. She reported on the MAGERT position: it is acceptable to drop position 24. Unfortunately, LC cannot drop byte 25 because it is used for duplicate detection and resolution. It must remain as is. Paul moved to drop byte 24 but not byte 25. Larry provided a second to the motion. Seven members voted in favor of the motion and no members voted against it. PROPOSAL 96-8/009: Canada proposes to add Field 009 to carry detailed data on Cartographic Material. It would be a detailed physical description fixed field. Unfortunately, part of the data defined in the Canadian 009 is currently already defined in other fixed fields, as detailed in the Proposal. There was lots of discussion and explanations about specific codes. For instance, what is a landset image and how are they handled? Mary Larsgaard explained that they are collected as digital data and then printed out. UCSB has several million landsets coded as 007/04 "a". There was discussion about the separate supplement code (008/25 "f") and bound with code (008/25 "g"). Betsy agreed that they don't fit well, but they do occur and are needed. Paul felt that the 007/06 codes and terminology were confusing. In general, many felt that more analysis and information was needed. Definitions could be improved upon, and it might be better to hold back on a vote until the 007 issue could be fleshed out in more detail. Priscilla asked for a straw vote on whether to vote now or postpone. It was about even. Priscilla and Sally decided to table the Proposal for the present. ***************************************************************** Monday, July 8, 1996 ***************************************************************** Priscilla suggested that the Committee begin at the top and work in order, skipping those proposals reviewed the day before. She reminded everyone that there was no pressure to pass something at this meeting, if more information would be helpful to the decision-making process. PROPOSAL 96-8/LDR-17: Sally McCallum gave a brief introduction into the reasons for and against a new encoding level code. Young-Hee explained why Option 1 (adding value 3) is preferred by Canada. The abbreviated level is lower than minimal level because no series statement is included. Catalogers need this value as it provides a signal to add information to bring the record up to the minimum level. She went on to say that there is tremendous pressure to catalog more with fewer resources, and the abbreviated level code is a useful mechanism. Originally it was used for pamphlets and ephemera but now it is used for all sorts of materials. She emphasized that Option 1 is really preferred. John Attig confirmed with Young-Hee that Canada is willing to convert some records at the end of the harmonization process; in fact, the cost to NLC and Canadian libraries will be very high based upon what has already been agreed upon. Priscilla asked what determines a USMARC encoding level; she pointed out that OCLC has ENCL "i" and other non-USMARC codes. John Attig replied that it depends upon whether or not a code is useful beyond one utility or a limited user community. Young-Hee and others could see some usefulness for value 3 beyond the Canadian community. Robin asked for a quick summary of the differences between abbreviated, partial, and minimal. Young-Hee provided the following details: --Preliminary is used when an update of the record is expected. --Partial is used when distributing LC records to Canadian libraries. --Abbreviated is lower than minimal. NLC will not update ever but other libraries may do so. Betsy Mangan stated that the LC Map Division would find it very helpful to have the value 3 added to the Encoding Level byte. Paul Weiss saw a use in managing a cataloging workflow. He would assign higher-level staff to value 3 records, since authority control work may be required on these records. Sally asked Rich Greene if OCLC has a similar code already; unfortunately not, since most of the OCLC codes are variations on the current USMARC codes. Catherine Gerhart agreed with Paul about workflow management. Priscilla asked if anyone had a problem with adding Value 3, if it wasn't for the system impact as detailed in the Proposal. For the most part, no, except that different standards might define Abbreviated differently and this could cause a problem when records are exchanged. Rich added that a cataloger may need to look at the 042 to see whose definition of "abbreviated" applies to the cataloging record. Sally spoke about the need to reconcile definitions for minimal cataloging. Bonnie Dede spoke up for convenient access to each definition, perhaps in a USMARC appendix. Others felt it was more appropriate for the USMARC codes to be generic, and that utility or system documentation should be specific. Priscilla agreed with Bonnie that it was easier to manage abbreviated records if they do have a unique encoding level. In that sense, most people were in favor of Option 1. Priscilla Caplan then asked the Committee to discuss the issues raised in the impact statements. Rich Greene reported that it would be a very painful process for OCLC to add any new encoding levels because the byte is used extensively throughout the software. The work for value 4 (Core Level) is not yet done; in fact, if value 3 is passed, OCLC would probably do both together. Young-Hee asked if OCLC could drop "3" records until the new encoding level was programmed. Rich confirmed that this could be done without much effort. Young-Hee reported that NLC is even now releasing "3" records; it is unknown how LC processes them. Perhaps they are dumped and not distributed to USMARC record subscribers. Sally asked if OCLC would want libraries to input "3" records, if the value was passed. Yes, Rich said, since OCLC libraries might also find the code to be useful for workflow management. Priscilla asked other systems for comments. Joe Altimus reported that RLG was comfortable with adding value 3 to byte 17 of the Leader. Marti Scheel reported that the National Library of Medicine supports the change. Michael Johnson, representing MicroLif, felt that some if not many system vendors would find the additional code to be useful. Priscilla summed up by saying that people are generally in favor. Adding the new code is hard for OCLC to do, and more information is needed from other systems who might have some difficulties. PROPOSAL 96-8/007-02: The usefulness of the codes for facsimile, original, and reproduction has been questioned now that more and more bib records are merged. Also the values have been difficult to distinguish. Sally said that U.S. practice has been in error in the past, whereas the Canadians did not make the same mistake. Young-Hee said that Canada would accept either option. Rich asked for clarification on the end result of the two options. Basically, byte 02 would end up with a different definition (blank versus obsolete). Paul supported Option 2; he trusts catalogers to stop using an obsolete code; there is no system effort to change existing records with any obsolete value. After more discussion, Paul moved to accept Option 2. Robin provided a second to his motion. There were five votes in favor of the motion and no votes against it. PROPOSAL 96-8/008-39: Sally explained that NLC would like a unique Cataloging Source value so that NLC records would not be considered to be LC records. Three options are given in the Proposal: --Option 1: "f" for NLC --Option 2: result is a simplified set of values --Option 3: more generic set of codes Young-Hee reported that Canada prefers Option 1, but could live with Option 2. Option 3 involves too many changes. Marti Scheel reported that NLM has used the "b" code consistently since 1993 and does want to retain it. Michael Johnson wondered if the list of national libraries will keep expanding in the future. Paul suggested making the byte obsolete, since the codes have been used inconsistently in the past and usefulness is minimal at this point. Rich Green reported that OCLC uses this byte a lot in its software. Adding codes for new national libraries requires some work, but not as much than if the code was made obsolete. Diane Hillmann agreed with Paul. Bill Jones asked if Option 3 implies you look elsewhere for the information. Sally replied in the affirmative. Robin said Harvard would find Option 3 inconvenient since the byte is used for system processing in combination with the Encoding level byte. Gary Smith confirmed that it is not so hard for OCLC to add new codes to byte 39, whereas it is harder to change existing codes or make the byte obsolete. To others, Option 3 seemed more rationale as more and more countries begin to use USMARC (or IMARC as discussed in the Business Meeting). Priscilla asked how catalogers use the code in their work. Catherine said that it depends upon the system in use and what fields display on a given screen; a cataloger may need to look at the 040 field. The cataloging source code is very useful during retrospective conversion of records. Gary Smith liked Paul's suggestion to make the code obsolete. Sherman Clarke thought Option 3 would be OK for first-level sorting of incoming materials. Priscilla concluded that the Committee could not settle on one of the three options at this meeting, but that the discussion had been useful. PROPOSAL 96-8/008-24-27: Sally explained that NLC proposes adding codes for Theses and Treaties to the Nature of Content byte. Young-Hee had no comments. Bonnie requested two editorial changes: The definition for the "m" code should be "theses" and the definition for the "j" code should be "patent documents." Rhonda Lawrence pointed out that most U.S. treaty sets are serial records and that some universities have "best thesis" series. Since the Nature of Content byte has been in alignment for the most part with bytes 24-27 (Serials), this raised the issue of adding the same codes to the Serials bytes. Randy and others pointed out some ways that the two lists have not been in perfect alignment: some codes are obsolete in one list but not in another list, "z" is defined inconsistently, etc. Larry Woods moved that codes "m" and "z" be approved for both books and serials, with the editorial changes suggested by Bonnie. Paul Weiss made a second. Six MARBI members voted in favor of the motion and no members voted against it. PROPOSAL 96-8/008-33: Sally introduced the Proposal by saying that USMARC has in past had an on/off byte for fiction, whereas CanMARC has had the same on/off code as well as specific genre codes for the Fiction byte. The proposal is to make values 0 and 1 obsolete, and to add various "alpha" codes for various genre. An impact comment pointed out that mapping from the old codes to the new codes would be impossible by machine. Young-Hee pointed out that the current list of Canadian codes is closer to what is used in UKMARC and UNIMARC. Larry Woods and John Attig expressed concern about the "scope" of the byte; to what extent should genre codes be assigned here? Larry felt there was too much mixing of apples and oranges. Bonnie Dede agreed that the new codes clouded the Fiction description of the byte too much. She asked if the Nature of Contents code might not be a better place for the Canadian codes. Sally discussed the trade-off between fixed field codes versus variable field codes or phrases. True codes seem to work better in international usage. A British attendee reported that the UKMARC codes for various genre in this byte have been assigned to records in the BLAISE system and are used for UKMARC users for retrieval. Sally asked if there was any known difficulty with the overlap in definitions. Perhaps a little during training but it generally doesn't seem to be a problem. Short stories is viewed to be a form of fiction, but the unique code is useful in a practical sense. Rich Greene expressed mix feelings, although he understands and wants to support uses outside of the U.S. He believes that the Brazilians see this byte, as currently defined, as a problem in their use of USMARC. Paul remarked that the ALCTS SAC Form/Genre Subcommittee worked on a coded list, but decided that text or phrases worked better in the end. He suggested retaining "0" (Not Fiction) and "1" (Fiction). He expressed concern about code "m" (Miscellaneous) and preferred a definition of (Multiple Forms) for "m" with another code for (Other). Catherine suggested using "u" for (Unknown) instead. Someone else preferred using the fill character rather than a pound sign for (Non- ficton). The Proposal was amended as follows: Retain codes "0" and "1" and add "type not further specified" to each definition. Make "#" ("Non-fiction") obsolete in CAN/MARC so that Canada will not have to convert records if not desired. Change the definition for "f" to "Novels" (not Non-Fiction) and value "m" from "Miscellaneous info" to "Mixed forms." Paul moved to pass the amended Proposal. There were seven votes in favor of the motion and no votes against it. PROPOSAL 96-8/008-18-19: Canada requests that three codes be added to the list of Form of Composition bytes; this is Option 1. Option 2 proposes making these bytes obsolete. The three proposed codes do not present problems for those already coding and the Music Library Association is in favor of Option 1. Paul questioned the "sd" (Square dance music) code, since square dance music is not a form of composition even if the call is embedded. Karen Little agreed with Paul. The music community realizes that the field is a mess: the definitions are inconsistent and the code list is incomplete. However, overhauling the whole list is not feasible and allowing new codes for those who want them is the best course at this time. Larry moved to approve Option 1. Sandra Lindberg provided a second. There were seven votes in favor of the motion and no votes against it. PROPOSAL 96-8/008-23-27 There is a typo in the Proposal; there are five bytes in this field, so attendees were asked to change "24-27" to "23-27." Sally explained that the bytes are defined for accompanying matter for visual materials, especially archival motion pictures. LC does not use currently. It is known that the UCLA Film Archives does not use. It is unknown whether the field is used in other USMARC systems. CanMARC defined byte 23 for "Form of Item" and left only four bytes (24-27) for "Accompanying Materials." Sally summarized the three options in the Proposal: change American values to Canadian values, add Canadian values but leave American values alone, and make the field obsolete. Young-Hee reported that Canada did not consider the obsolete option, and she would like to take this back for discussion. Maureen Killeen reported that, in a sample of 20,000 Canadian records, only 6% had these bytes coded. Paul thinks that the 006 and other places in the MARC record are a better place for Accompanying Materials codes. Rich Greene reported that OCLC staff prefer the obsolete option; the other options would be hard to do. Gary Smith added that since you can't tell when something is coded by a cataloger, you can't switch your list of codes mid-stream. Priscilla asked if others agreed that Option 1 is not viable. Yes, there was agreement. Young-Hee asked the same question about Option 2, because Canada didn't think it was viable. John Attig suggested that there not be a vote today. Let the bibliographic community discuss Option 3 is more depth. Sally agreed to re-write the proposal, favoring Option 3. PROPOSAL 96-8/016: CanMARC has field 016 defined for the NLC record control number; it is comparable to field 010 used for the LC record control number. It is the citation number in Canadiana. This Proposal has three options: #1 Add the 016 field for the NLC Control Number. #2 Change field 010 from LC Control Number to National Bibliographic Agency Number and add $2 for the code or name of the agency. #3 Make both 010 and 016 obsolete. Use field 015 for National Bibliographic Agency Control Number, with undefined indicators and subfields $a and $z. Young-Hee reported that Canadian systems process this number extensively. Field 015 is also used a lot and the numbers vary, so she believes that it would not be good to go with Option #3. Priscilla proposed another option: use 016 with a $2 to show bibliographic agency. She thought generalizing a new USMARC field would be best in the long run. John likened this idea to the 035 field, and Young-Hee agreed. Another stated the 035 is a junkyard. Paul felt the easiest solution is to define the 016 and have a blank indicator for NLC and other indicators for other agencies. A visitor from the United Kingdom reported that the 016 is used in UKMARC to show authority control in a bib record. Priscilla pointed out that there is no option that won't require conversion by someone. Paul moved to accept Priscilla's suggestion. Define 016 as a national library control number. Assign the blank indicator to NLC. Define the 7 indicator to refer to a $2. Define subfields $a, $z, and $2. Use the NUC symbol in the $2. There was more discussion about the name of the 016, and there was some concern about the 015 field. It was agreed to develop this new option in a second Proposal. Paul withdrew his motion. PROPOSAL 96-8/028-Ind 2: Sally introduced this Proposal and stated that the problems here are very difficult. CanMARC overlaps with and reverses values from USMARC. Another complicating factor is the fact that some libraries still use these USMARC codes for printing added entry cards. The Proposal gives two options: #1 Redefine into in/off display of the note. #2 Retain the USMARC values unchanged. Karen Little reported that the Music Library Association wants to retain the ability to distinguish and maintain an 028 as an added entry. Marti Scheel reported that NLM supports this position, and prefers Option 2. Kathy Bales suggested adding two new codes with the Canadian definitions. Young-Hee pointed out that this defeats the goal of harmonization, and John pointed out that record conversion would still be necessary. Paul reported that his system uses this indicator for indexing and display. John supported the MLA position, that libraries printing cards should not be forced into printing unneeded cards and then tossing them. Young- Hee stated that it would be an expensive conversion for Canada, if the USMARC codes are imposed on them. Priscilla didn't think the conversion would be so difficult. Rich was concerned that card production would be a mess if the meaning of the codes was reversed in some records but not all. Maureen reported that ISM's position is neutral on the two options; ISM does use this indicator for printed products but also USMARC records are allowed in the ISM database now. Priscilla said that perhaps the only way to be fair here is to make the field obsolete and start all over again. The consensus was to revisit this Proposal at the Midwinter meeting. PROPOSAL 96-8/082-Ind 2: In the Dewey Decimal Number field, it is proposed to add one value and adjust the existing ones as follows: 0 Assigned by LC Dewey agency 1 Assigned by other national Dewey agency 4 Assigned by other agency. Sally reported that the LC Dewey Office likes this proposal. To be able to distinguish who has assigned a Dewey number is useful in several ways: for teaching in the field, for promoting consistency in the assignment of numbers, and for automatically generating mappings between LCSH and Dewey numbers. John was not sure that the proposed Indicators of 1 and 4 would be helpful; Rich reported that it did help in a large database of shared records when multiple agencies assign Dewey numbers. Young-Hee suggested that an alternative would be to let a blank mean "Not Specified." If this were done, Canada would not have to convert records. Sally said that blank had the "Unspecified" before and that it is now "Obsolete." This could be withdrawn. Priscilla asked if there were any objections. No. Any more discussion? Sherman asked about Dewey-like numbers and Young-Hee explained they were handled in the 082 field. Paul moved to accept the Proposal and to re-activate the blank code. Another member provided a second. There were seven votes in favor and no votes against this motion. PROPOSAL 96-8/085 Canada requests a new field to carry the CODOC shelving number that is applied to Canadian federal and provincial government documents in Canada. It is a variable-length code consisting of alphanumeric characters. CanMARC currently uses the tag 088 for the CODOC number and is willing to change since it conflicts with the field 088 (Report Number) in USMARC. Sally explained that the CODOC number is like the government numbers in field 086 but it is assigned by non- government agencies and is a local schema. Two options are presented: #1 Use field 085. #2 Use field 084. Young-Hee reported that Option 1 is preferred because conversion will be much easier. The 086 was considered by the Canadian community but there are some subtle differences. The CODOC number is generated by the library cataloger, instead of being assigned by a government agency like the SUDOCS number is assigned by the Superintendent of Documents. Sally said it would be possible to redefine the 086 definition, if people wanted to consider the 086 as a third option. Young-Hee thought that indexing could be a problem in systems, if the CODOC number was added to the 086 field. Diane Hillmann liked the 086 option, since the number sounded to her like a SUDOCS number. Robin and Priscilla disagreed; they thought it was closer to an 084 number. Diane Hillmann and others expressed concern regarding Option #1, from the point of view that it is not a good idea to assign a new field tag in a limited circumstance. It was agreed that more information is needed on the assignment of CODOC numbers by the Univ of Guelph cooperative project. Sally suggested that this Proposal be postponed until Midwinter while she and others did some more analysis. PROPOSAL 96-8/9XX: Sally introduced the topic of CanMARC 9xx fields and referred to the remarks on Saturday by Ingrid Parent. These are not really local fields. They are legitimate names, uniform titles, and cross-references in equivalent French and English forms. The data logically resides in authority records, but is needed in bibliographic records if a local system has no authority control module. Unfortunately, the CanMARC 9xx fields conflict with USMARC local fields. Young-Hee explained that Canada is a country with two official languages, and that this condition creates specific bibliographic needs. Unilingual publications are cataloged in the language of the item but include subject headings in both English and French. Two separate records are created for bilingual publications with the English language record carrying English subject headings and the French language record carrying the French subject headings. This is a very sensitive subject politically. Years ago, Canada selected the 9xx fields for equivalent headings in order to not conflict with USMARC. Paul asked why cross-references are included, since these are not language driven. Young-Hee said that it is the legacy of the past; it may no longer be done, but there are still records with 9xx cross-references. Priscilla reviewed use of 9xx fields in USMARC. If a single library or system defines, this is OK. If a group of libraries agree to use locally, this is OK. She suggested that we look at the Canadian practice in the same way. Information about the Canadian 9xx fields must be distributed, and each system would decide what to do with incoming 9xx fields. The question becomes how to document. Young-Hee had some concerns about this approach, because, in CanMARC, 9xx are not local fields. Also, who would be the maintenance agency for the documentation? Under Option 1, it would be USMARC. Under Option 2, it would be NLC. Paul felt that there is a better way to deal with translated data, and that treating the Canadian 9xx like local fields does violate USMARC principles. Bonnie agreed. The bibliographic community needs to gear up to better serve users with different primary languages. She suggested MARBI work on some longterm solutions for the bigger issue. John asked, if Option 1 (CanMARC 9xx documented in the official USMARC documentation) is selected, would Canada expect U.S. systems to load and process? Young-Hee said not at all. It would be published for the information of those receiving the equivalent language fields, but these fields could be dropped before loading. Michael Johnson observed that it would be impossible for vendors to support different local fields when distributing records. Young-Hee repeated her suggestion that the fields be deleted before loading into U.S. systems. Priscilla pointed out that Canada will continue to use 9xx fields. The practical issue at hand is how to handle the documentation. John agreed. Young-Hee emphasized that the 9xx fields are part of the CanMARC standard, and will become part of USMARC after harmonization is completed. Paul did not like Option 1 because, in his view, it would set an unfortunate precedent. He felt that the CanMARC use of 9xx fields is no different than OCLC's documentation on 9xx fields. Catherine said that as a cataloger she has been fighting for authority control modules in local systems for a long time. She did not want to open the door to having authoritative name fields in bibliographic records, even if these are equivalent fields only. John said that Canada is now paying a price for having solved an important problem using the tools available at the time. It is now time to look to the authority format to control multi-lingual terminology. Priscilla suggested that it would be good to follow the Internet Metadata approach. Establish a 9xx registration process, so that more information is available in a central place. Sally agreed completely and suggested that it would not be hard to publish registered 9xx definitions on the USMARC home page. Sally and Priscilla asked if there were any further comments. No. It was agreed to defer the rest of the CanMARC authority proposals until the Midwinter Meeting in February 1997. Discussion is encouraged on the USMARC list in the meantime.