Priscilla Caplan LITA Univ of Chicago James Crooks RASD UC Irvine Shelby Harkin ALCTS Univ of North Dakota William Jones LITA New York Univ Jacqueline Riley RASD Univ of Cinncinnati Gary Strawn ALCTS Northwestern Univ Flo Wilson ALCTS Vanderbilt Univ Larry Woods LITA Univ of Iowa MARBI Interns: Josephine Crawford ALCTS Univ of Wisconsin Paul Weiss ALCTS Nat. Lib of Medicine Representatives and Liaisons: John Attig OLAG Penn State Univ Jui-Wen Chang RLG RLG John Espley AVIAC VTLS Catherine Gerhart CC:DA Univ of Washington David Goldberg NAL NAL Richard Greene OCLC OCLC Robyn Greenlund MicroLIF Data Trek, Inc. Rebecca Guenther LC LC, Net Dev/MSO Karen Little MLA Univ of Louisville Sally McCallum LC LC, Net Dev/MSO Susan Moore MAGERT Univ of Arizona Patricia D. Sesha ARLIS Frick Art Ref Lib Marti Scheel NLM NLM Stuart Spore AALL NYU Law Arlene Taylor SAC SAC Rutherford Witthus SAA Univ of Colorado Young-Hee Queinnec NLC NLC Other attendees: Joan Aliprand Research Libraries Group Hiroko Aikawa Cleveland Public Library Virginia Berringer University of Akron (Ohio) Diane Boehr Castabile Associates Candy Bogar DRA Roger Brisson Pennsylvania State University Ann Case H.W. Wilson Co. Sherman Clarke Amon Carter Museum Marilyn Craig University of Houston Carroll Nelson Davis Columbia University Emily Epstein University of Colorado at Boulder Ching Eue Houston Public Library Lori Franz University of Michigan Linda Gabel OCLC Bob Hall Massachusetts Institute of Technology Susan Hayes New York University Lynne Haymon University of California, Riverside Donna Heist University of Iowa Renee Helinas Getty Trust, Bibl of History or Art Beth Helsel Clemson University Harrietta Hemmasi Rutgers University Elaine Henjam Florida Center for Library Automation Diane Hillmann Cornell Law Library Gail Hodge NASA Center for Aerospace Information Elizabeth Geery Holmes Roger Williams University Law Library Lee Ireland DRA Bruce Johnson Library of Congress Coby Johnson University of Montana Sherry Kelley UCLA Ken King University Microfilms Inc. Jim Kuhn University of Akron Mary Charles Lasater Vanderbilt University Sara Shatford Layne UCLA Sandra Lindberg (not supplied) Dorothy Lourdou New York Public Library Winslow Lundy Bryn Mawr College Joan Lussky Technology Horizons Peggy Mahan Jenkins Law Library Bethany Mendenhall Getty Center Carolyn Miller University of Michigan Catherine Nelson University of California, Santa Barbara M. Diane Raines Retro Link Associates Debra Shapiro Chicago Historical Society Margaret Shen Cleveland Public Library Sushila Sheih Manchester College Jackie Shieh University of Virginia Judy Silverman Canadian Centre for Architecture Patricia Sistra Frick Art Reference Library Holly Schmidt Blackwell North America Judy Silverman Canadian Centre for Architecture Sally Sinn National Library of Medicine Ann Sitkin Harvard Law School Library Gary Smith OCLC Steven J. Squires University of North Carolina Daniel Starr Museum of Modern Art Lennie Stovel RLG Jennifer E. Sweda University of Pennsylvania Louise Swold Cuyahoga County Public Library Millie Della Terza Houghton Library, Harvard Cecilia Tittemore Dartmouth College Library Belinda Urquiza Library of Congress Bob Warwick Rutgers University Libraries Robin Wendler Harvard University Gregory Wool Iowa State University MARBI Recorder: Josephine Crawford University of Wisconsin ******************************************************************** February 4, 1995 ******************************************************************** Chair Priscilla Caplan welcomed members and guests to the first MARBI meeting of the 1995 Midwinter Conference. After introductions around the table and a review of the day's agenda, the meeting moved quickly to the first proposal of the day. ----------------------------------------------------------------- PROPOSAL No. 94-10 : "Encoding of Patent Information in the USMARC Bibliographic Format" This proposal was introduced by Sally McCallum. The original request had come from the LC National Translations Center which is now closed. Patent depository libraries have been set up in each state, and the NASA Library has a strong collection as well. If the proposal is passed, Sally is looking for a few interested libraries, in order to work with them and hammer out problem areas. The descriptive data elements for patents have been standardized at both the national and international level using an INID numbering scheme (INID stands for Internationally-agreed Numbers for the Identification of Data). Many of these data elements can be mapped to one or more USMARC fields, but some cannot be mapped at all. In order to accommodate all of the bibliographically significant data elements from the patent document standards, some new USMARC data elements are proposed in Section 6 of the proposal. John Attig opened the discussion by questioning what is being described. Sally responded by stating that the "bibliographic item" in a patent collection is a patent description with one or more cover sheets (e.g. both a US and a foreign application). The cover sheet(s) describe the patent document and its processing state, such as filing and publication data. The patent document itself is a description of the patent. She went on to state that the purpose of the proposal is for catalogers to be able to catalog patent documents added to their collections, not to convert specialized patent records maintained by the U.S. Patent Office. John Attig questioned why the proposed 013 field is repeatable. Would this be for each cover sheet? Answer: The field is repeated when more than one number is assigned to the patent. Rich Greene suggested changing $n (Number) to $a to be more consistent with the other number fields. Sally suggested striking two 013 codes ($c Kind of document code and $h Country code) because they are already found in the 008 field. Someone else questioned separate 013 subfields for both $g country name and $h country code. Are both needed? Why not just the code with a display constant? Paul Weiss believes that it is important to avoid new content designation when possible. For example, the kind of document code might be handled by a 655 field. Also, should the $k be linked with the number? If not, it would be better in the 655, that is, if record level. Someone suggested looking at various alternatives. For instance, perhaps a subfield could contain either a name or the matching code. Has a standard code list been completed? Can documents have either a name or a code? Is it worth having both in the USMARC record? Perhaps require one of the two? Subfield $g (Country Name) should be parallel with $h (Country code). Bill Jones suggested that the last sentence of the $h description be added to the $g description. Sally agreed to look further into the $c (Kind of Document code) by consulting with patent library staff. It was questioned if the various dates ($d, $e, and $f) are important for tracking purposes. The dates relate to the document as a whole and are very important legally. The description on page 12 needs more clarification to distinguish the three date subfields. Sally will check into whether the $d is the date published, the date of receipt, or just whatever date is available. Larry Woods asked if all the dates remain on the cover sheet; yes, Sally thought. It was suggested that a single subfield might work for all three dates where each date has a qualifier in parentheses. Priscilla suggested following the ISBN form where the $d has the date and the $f has the kind of date or qualifier. From a programming point of view (both display and indexing), Karen Coyle prefers an explicit subfield code for each type of date. Randy and Sally want to look into using the 008 field for the language. They plan to look at various examples first, but they may delete the $l from the 013 field if they think the 008 fixed field will suffice. Young-Hee Queinnec would like to see the note-like data be more consistent with other notes in USMARC. For instance, perhaps $i (Information Phrase) could be placed in another standard field, like 506, 540, or 037. There is some discussion about merging 506 and 540 anyway. There was consensus in the group to use the appropriate 5xx note field. Will need to look into linking entry fields and a linking entry complexity note. Page 8, INID 71 should read "applicant" and not the plural form. Priscilla pointed out that the $c description on page 11 contains a requirement for left justification and this is not normal for variable fields. The structure of the number in $n is YY-AANNNNN where YY contains the last two digits of the calendar year. Sally expects that we will need a four digit year code at the turn of the century, and suggests changing the form now. Gary Strawn wonders if the cataloger shouldn't record what is present on the item itself, two- or four-digit as appropriate. It was agreed to write a letter to NISO to ask them to change to all four digits for the year. Paul Weiss asked if catalogers would normally follow AACR2? Where possible/applicable. Will a 300 field be included? Will names be under authority control? Will any punctuation rules be followed? Paul sees a problem in supplying 7xx indicators. Sally responded that the cataloger can use any USMARC tagging that is appropriate, and that the same issues raised by Paul relate to any records created where AACR2 is not followed (such as GILS records). A motion was made by Gary Strawn for LC to rework the proposal for the June meeting, taking into account the various comments and questions. The motion was seconded by Bill Jones and Shelby Harkin. The motion was passed. ----------------------------------------------------------------- PROPOSAL 95-1 : "Changes to Field 856 (Electronic Location and Access) in the USMARC Bibliographic Format" This proposal was introduced by Rebecca Guenther. The title should be changed to "Bibliographic/Holdings" Format. There are several specific purposes in these changes: -- want to align $2 with the Internet Uniform Resource Location (URL) draft standard. (Lenny Stovel pointed out that there have been several recent changes and the standard now exists as an Internet Request for Comments (RFC) document.) -- add hours of host computer. -- add ability to handle non-Internet resources ($b, $r, $v) -- add value "3" to first indicator for dial-up. Priscilla suggested a discussion strategy: first, talk about proposed changes; second, talk about what we don't like about this field. Paul Weiss suggested standardizing the form of the phone number for automatic logon (area code + 7-digit number). What about local dialing needs, where a 1 or 9 is needed to get an outside line? What about in-house extensions? For international calls, a country code is also needed. Is there any NISO activity to standardize phone numbers? Agreed to include this suggestion in the letter to be sent to NISO in near future. Bethany Mendenhall suggested using subfields for country code, area/city code, etc. But not enough subfields are available unfortunately. A consensus was reached to begin with the country code, followed by the other numbers in logical order. A hyphen will be used as a separator, and an "X" will be used before the extension. Example: $b 1-213-1234567-X1234 Rebecca agreed to look at the Community Information Format to see about bringing those phone numbers into consistency with this form. In regards to $j (baud rate), highest to lowest syntax was thought to be confusing. It was agreed to change $j to "Lowest BPS - Highest BPS." In regards to $r (parity), it was agreed to specify all possibilities in the description: O (Odd), E (Even), N (None), S (Space), and M (Mark). Karen Coyle asked if the $b (access number) could be repeatable? Because some systems have more than one number. If the different access numbers must be correlated to a specific baud rate, than a new 856 should be used for each. Agree to repeat the $b only if all the information in the rest of the field applies. Rebecca is writing guidelines for LC staff, and she plans to post them to the bibliographic community. In regards to $u (Uniform Resource Locator), Bill Jones asked if the URL label should be used. Rebecca responded that the presence of the label complicated access to their Web server. She is ready to remove it from the example since it is redundant anyway. It was requested for instructions to be added for when it is possible to repeat the $u; it can be repeated if the rest of the data in the field applies. Rebecca will also add more examples. What will happen when other developing Internet standards like Universal Resource Name (URN) and Universal Resource Identifier (URI) are ready? This question will be quite complex, involves version issues, total remake of MARC, etc. A motion was made by Gary Strawn to accept the proposal as amended. Another member seconded the motion. The motion was passed. ----------------------------------------------------------------- PROPOSAL 95-5 : "Addition of Subfield $b to Field 154 (General Explanatory Index Term) in the USMARC Classification Format" This proposal was also introduced by Rebecca. LC staff didn't think a 154 field $b was needed in the original creation of the classification format. But, they have now found a need in limited situations during LC classification conversion. These situations are described in the proposal. It was not clear from the proposal if multiple occurrences of subfield $b are to imply successive levels of the index term, or as alternatives at the same level. The idea of 95-5 is to align classification field 154 with field 753. Paul Weiss pointed out that in field 753, if you have a $u you have to also record a $d, which is contradicted by the example in 95-5. After brief discussion, it was agreed to reword the description of the field so that $d is NOT required. Gary Strawn made a motion to pass the proposal as amended. Flo Wilson seconded the motion. The motion passed. ----------------------------------------------------------------- DISCUSSION PAPER NO. 80 : "Defining a Component Item Entry Field in the USMARC Bibliographic Format" This discussion paper was introduced by Lennie Stovel of Research Libraries Group. Two RLIN libraries currently use a local field to contain information about items that are constituent parts or elements of an item being cataloged. There are two separate aspects to the proposal: description of the constituent parts and linking to an electronic location. The discussion paper suggests the creation of a new linking entry field, 774, for the description. The 856 field could be used for electronic location and access, linked to the 774 by a $8. Alternatives are described in sections 6 and 7 of the paper. John Attig suggests another possibility : using the $8 in various repeating fields for a more complete description. He worries about large numbers of 774 fields, creating a monster record. However, Lennie believes that RLIN hasn't hit any record size limitations (Sherman added though he knows of one case where the size of the record caused an artificial separation). Catalogers do tend to separate large collections into logical subsets. Rich Greene mentioned the multi-level description rule in AACR2 13.6A. The cataloger can create sets of paragraphs, each level independent. This is not well implemented in USMARC. Robin Wendler reported that Harvard is using a local field in the same way as the other RLIN libraries. It seems appropriate for extensive content information. John prefers separate records with links. However, if a separate title does not exist, there is no 245 for the separate record. Diane Hillmann believes that a range of options are needed due to the real-life variations present in materials and workload and system factors. Priscilla Caplan noted that three cataloging options have existed to date (added entries, content notes, and separate analytic records); is a fourth option necessary? Sally McCallum stated that the analytic approach is thought to be the standard approach. Paul Weiss suggested yet another alternative: use an 844 field in the holdings format. May need more subfields for this situation. One problem though is putting bib information in the holdings record. Supporting Diane Hillmann's point of view, Robin Wendler thinks that the 789 approach works well for certain types of materials (e.g. drawings in a set). Several projects independently came up with the same solution, an argument for its usefulness. Lennie pointed out that RLIN did not attempt to standardize all the field's subfields, so that there could be local definition appropriate for the local collection. Susan Moore, representing MAGERT, stated that topographical sheets have been done as analytics. Due to the large number, if done as multiple 774 fields, would fill up the record real quick. John Attig pointed out that a 774 field, in the linking block of fields, is not really a linking field, it is a description, so another tag would be better. Sally pointed out that linking fields can link to a "virtual record." Although, Sally wondered if linking fields shouldn't have a URL for the object they are pointing to? The paper sort of recommends putting the URL in an 856, rather than use the 774 as in the RLIN applications, in order to make use of an existing MARC field. John doesn't think that we should be too restrictive in placing URLs in 856 fields only. It was agreed to flesh out these issues in a second discussion paper. "Linking" needs to be better defined in this context. Paul Weiss volunteered to work with LC staff on the description issues. John Attig feels that the component part definition in USMARC needs to be changed and should be looked at as part of the second discussion paper. ----------------------------------------------------------------- BUSINESS MEETING : Larry Woods announced that a Unicode meeting would take place on Sunday night at 7pm in the RLG Suite at the Marriott. Have much work to do. Sally McCallum announced that a new USMARC specifications document has been published; now includes Arabic and Extended Arabic. Holdings Format Update #2 has been published. Are working on Bibliographic Update #1 to be published in August; it will include the 1995 Format Integration changes as well as the June '94 - Jan '95 approved changes. Sally asked for advice on republishing the Concise edition. It is really out of date now. Do people use it? Several copies had been brought to the conference! Perhaps an electronic concise is the way to go, like the Format Integration Concise which was just posted on Marvel. What about the Relator list? It has much description; it is not yet on Marvel. There is agreement that an electronic version is needed. Recent questions have come up regarding language codes. NISO has come up with a new list, with all USMARC codes plus some new ones. Sally intends to redo the USMARC list to match NISO's, and also add some other needed Canadian codes. The ISO list of language codes is in progress, but this will be a very slow process internationally. It is mostly the same as USMARC, except for about 20 codes which are different based upon political compromise. It was announced that Marvel now has WAIS indexing, and the menu picks were recently reorganized. Some members would find it helpful if the USMARC list was archived more often than monthly. The Phase I Format Integration training has been completed by LC trainers. The records will be distributed next month. Sally announced that the year 1998 has a special significance, because LC would have been assigning LC card numbers for 100 years! Some system changeover will be required, so that 1898/1998 and later years do not conflict. Sally is writing up this problem which will be posted on the USMARC list as DP84; she requested input about the impact on various systems. Priscilla Caplan announced that a special email list exists of voting MARBI members only. The first electronic vote occurred last fall on Proposal 94-17a (New Proposed Field 786: Source of Data Entry). Priscilla asked if the vote was allright and if more electronic votes should occur in the future? She reported that ALA may come up with a standard policy on electronic votes in the future. Shelby believes that an electronic vote only works on straightforward issues. Flo agrees; she finds there is much value in meeting discussion for complex votes. ***************************************************************** February 5, 1995 ***************************************************************** The second meeting was opened by MARBI chair Priscilla Caplan. About seventy people attended. Priscilla asked Rebecca Guenther of LC to introduce the first proposal. ----------------------------------------------------------------- PROPOSAL 95-3 : "Addition of Subfield $6 (Linkage), Field 066 (Character Sets Present), and Field 880 (Alternate Graphic Representation) to the USMARC Holdings Format." The purpose of 95-3 is to allow the use of non-Roman data in holdings fields as needed by libraries; known examples include captions, notes, parts of call numbers, and the name of the holdings institution. The Research Libraries Group has already seen situations requiring use of the $6 for this purpose. The proposal seeks to define subfield $6 in all variable fields of the Holdings Format except 010, 027, and 841 (because the matching fields for these tags in the Bibliographic Format do not contain $6). The proposal also defines Field 066 (Character Sets Present) and Field 880 (Alternate Graphic Representation). John Attig pointed out that, when the holdings format was originally approved, it was explicitly stated that non-roman data would not be accommodated and that subfield $6 could be used for linkage. Rebecca explained why it seems appropriate to change this earlier decision now. The impetus of the proposal to change $6 to $8 in the holdings format was the later development of the classification format; also, the possibility of accommodating non-Roman data in the holdings format was mentioned in Proposal 91-3 when the Subfield $6 was changed to Subfield $8. Larry Woods pointed out that the Unicode implications are still unknown. There was discussion about the 022 field in the bibliographic format, with general agreement that the 022 should be added to the list of fields NOT containing a Subfield $6. Bill Jones made a motion to pass the proposal, with an amendment regarding the 022 field; a second was made. The motion passed. ----------------------------------------------------------------- PROPOSAL 95-2 : "Definition of Subfield $v for Form Subdivision in the USMARC Formats." This proposal was also briefly introduced by Rebecca Guenther. Currently form subdivisions are put into Subfield $x (General Subdivision), even though there are specific subfields for chronological ($y) and geographic ($z) subdivisions. The 1991 Airlie House Subject Subdivisions Conference recommended the creation of a separate subfield code for form subdivisions. As a follow-up, the ALA Subject Analysis Committee formally submitted a request to LC. There have been two discussion papers (DP 74 and DP 79) preceding this proposal. Arlene Taylor, representing the ALCTS Subject Analysis Committee, pointed out that several thesauri make a distinction between form and topic. Even though the implementation issues are complex, Arlene asked MARBI to support thesauri wishing to code a pre- coordinated subject string with a form subdivision identified as such. Historically, MARBI decision-making has followed a principle that the USMARC formats are responsible for "content designation" and not the actual content; Arlene requested that the same principle be applied today. Lynne El-Hoshy, representing LC Cataloging Policy and Support, summarized the situation from the point of view of her office. They have heard from several quarters that separate coding would be beneficial for pre-coordinated headings. Subfield $v and the 655 field are both helpful; they are not mutually exclusive; they complement one another. LC needs to develop a plan for the use of the 655 (the fiction project is using the 655 field now). It is recognized that there are times in cataloging other materials when a form heading is only appropriate as a subdivision and not as a 655. Lynne also reported that, as desirable as implementing the Subfield $v in 6xx fields is, LC cannot consider retrospective conversion until the LC system allows for global update. Josephine Crawford reviewed the results of the MARBI questionnaire distributed on the USMARC list last fall. There were 22 responses, representing a good mix of organizations. Questions 4-14 dealt with the issue of whether or not a "minimal implementation" was feasible; for some organizations, a minimal implementation involves only a small or reasonable amount of work; however, for LC Cataloging Service, even a minimal implementation is not insignificant. The questionnaire showed that a minimal implementation is feasible in many systems because a $v can be treated exactly like the $x in terms of indexing and display. A minimal implementation allows new cataloging records to be created with precise subfield coding. Only after retrospective conversion of existing files, does it make sense to take on the larger system task of treating the $v differently from the $x (in terms of indexing and display). Questions 15-23 dealt with the issue of retrospective conversion. Work needs to be undertaken to improve upon the conversion algorithm presented in the questionnaire, because respondents prefer the conversion to be as automatic as possible. Some excellent ideas were suggested in the remarks. Rich Greene opened the discussion by citing the general desirability of $v, and supported the view that a 655 field cannot act as a replacement for the $v. He believes that the USMARC format should support both, since both are valid cataloging mechanisms in use by the cataloging community. John Attig suggested that MARBI discuss the paper recently distributed by Arlene Taylor on the USMARC list, thereby taking the meeting right into the implementation issue. As chair, Priscilla Caplan decided that this discussion direction would be appropriate. Wearing her hat as a researcher who is interested in consistent and efficient OPAC displays, Arlene worked with a sample of 565 LC records from the OCLC database. She acted like a conversion program, comparing each subject heading against a draft list of dual-functioning headings. Then, she reviewed the titles for "aboutness" versus "form." Only 2% of the headings were wrong, an acceptable error rate in her view. She also discovered some specific patterns which a conversion program could handle. Priscilla asked for a showing of hands: is defining Subfield $v a good idea, apart from conversion? About half the people in the room raised their hands. Lennie Stovel said that she was convinced on the thesaurus side, but not yet on the display side. Sherman Clarke emphasized the importance of this proposal to the non-book bibliographic communities. Diane Hillmann sees it as the old chicken/egg issue: unless $v is defined, we cannot expect systems to program it in and make use of it in OPACs. Paul Weiss emphasized that the NLM can supply converted records for loading, since NLM's internal system maintains the distinction. Priscilla asked for arguments against defining Subfield $v. The discussion revolved around the cost/benefit issue and what priority implementation might take. John Espley maintained that vendors can do it, although users will need to pay for it ultimately. Robyn Greenlund sees it as the cost of doing business. Sally McCallum sees overwhelming support in this meeting, and concludes therefore that libraries will be willing to pay for both conversion and OPAC enhancements. Flo Wilson is not so sure that libraries will put implementing $v as a high priority. Robyn pointed out that library directors are not generally present at MARBI meetings. Larry Woods did not agree with Flo. It was acknowledged that the big impact on systems will hit when the utilities start supporting the $v. Bill Jones is particularly concerned with the implications of record migration, from utility to system to utility to system, etc. What do OCLC and RLG plan to do? Lenni Stovel spoke on RLG's behalf. RLG will add the Subfield $v to its system, if approved by MARBI. There are no plans to convert. Linda Gabel spoke on OCLC's behalf. OCLC staff are supportive of the principle behind the $v proposal. Unfortunately, current OCLC validation programs are based upon subfield coding; therefore, data validation will not work very well unless retrospective conversion is done. OCLC will have to write new software, if conversion is based upon an algorithm assuming the form subdivision is the last element in the field. Implementing $v for MeSH headings will be much less difficult; it is worthwhile to pass the proposal for MeSH and AAT. LCSH requires more analysis. Sally Sinn asked how the MARC documentation would address this? Perhaps use the phrase "as defined by the thesaurus being used." The ALCTS Subject Analysis Committee would like to see a coordinated, well-planned implementation, with perhaps MeSH and AAT being the lead thesauri, followed by LCSH on the schedule acceptable to LC and the utilities, organizations which must by necessity juggle many competing projects. Priscilla asked for a straw vote of voting members only, in order to gauge if discussion of specific proposal details was appropriate at this time. Six members voted in favor, and no members voted against in the straw vote. Priscilla asked everyone to turn to page 8 of the proposal to review the specific wording and the applicable fields. John Attig pointed out that something must be done with the 755 field. At one point in time, x00 $k was defined for form subheadings; thus, an editorial change will be needed in the documentation. In the authority format, Jui-wen Chang asked if the form subdivision be recorded in a $v or $a? To be consistent with current practice, it was agreed to use the $v. Paul Weiss suggested that a 185, 485, and 785 would be better than 183, 483, and 783. Sally and Rebecca will look into this. It was agreed to use a Bibliography example on page 11 in place of the Early Works to 1800 example. MeSH and AAT examples must also be added. Priscilla asked if anyone had any comments to make, for or against. John Espley stated that the vendors would very much appreciate more analysis on display issues. Arlene agreed that this is something that SAC will do, in addition to more work on conversion issues and implementation phasing. A query could also be sent to the USMARC list about display issues. Flo Wilson made a motion to accept Proposal 95-2 with the following amendments: add MeSH and AAT examples; use 185/485/585 fields in authority format unless already in use; change the Early Works example. Bill Jones seconded the motion. Six members voted yes, and one member voted no. John Attig urged the MARBI Committee to be concerned with the implementation issues of $v, perhaps through regular reports at the MARBI business meetings. Others at the meeting felt that the ALCTS Subject Analysis Committee would be the best place for overall coordination to take place (i.e. across all thesauri). ----------------------------------------------------------------- DISCUSSION PAPER #82 : "Merging Field 755 with Field 655 in the USMARC Bibliographic Format." Sally McCallum introduced the paper. Due to the fact that it has not been possible for catalogers to distinguish form/genre from physical characteristics in a consistent fashion, it is proposed to merge the 755 field with the 655 field. The SAC Subcommittee on the Nature and Use of Form Data could not come up with a definition which drew a sharp distinction, and this group supports the merge. The Library of Congress Graphical Materials Thesaurus gives instructions on how to use the 755 field, but the responsible people are happy to drop the 755 field and use the 655 field instead. There has been some email discussion on the topic, without a clear consensus emerging. Rutherford Witthus, representing the Society of American Archivists, spoke in favor of the merge. The archival community has discussed this issue in depth; most agree that it is too difficult to make the distinction for all materials. In addition, no systems index the 755 field separately from the 655 field (others in the audience agreed that they knew of no systems where the two fields are indexed separately). Therefore, there seems to be no real advantage in making the distinction between form/genre and physical characteristics. One member of the audience pointed out that there is a need to describe copy-specific information. This problem is solved since the 655 has a $5 with a copy-specific link. The next step is for a proposal to be generated from today's discussion. If this is done, what will be the name of the new 655 field? Would conversion be undertaken, and what are the issues? Gary Strawn moved for a proposal to be brought back to MARBI to make the 755 field obsolete in favor of the 655 field. His motion was seconded by Larry Woods. ----------------------------------------------------------------- DISCUSSION PAPER #83 : "USMARC Authority Records for Genre/Form Terms" Rebecca Guenther introduced the paper. There is a need to support the authority control of genre/form headings, by adding to the USMARC authority format. Two options are discussed in the paper. Option 1 would create X55 field for genre/form. Option 2 would use additional codes in the 008 field to indicate the heading's use as genre/form. Rebecca noted it is particularly desirable to provide MARC coding so that a reference structure across thesauri can be supported (i.e. LCSH now uses "Livres a clef" whereas the Guidelines on Subject Access to Fiction, Drama, etc. use "Romans a clef."). Rebecca also pointed out that the paper is NOT discussing authority record changes in support of occupation and function terms used in 656 and 657 fields in the bibliographic format. Arlene Taylor, representing the ALCTS Subject Analysis Committee, came out supporting Option 1 due to the fact that 655 fields are usually involved and a single database often has headings drawn from multiple thesauri. John Attig, speaking on behalf of the ACRL Rare Books community, believes there is a strong need to distribute thesauri as authority files. He has a weak preference for Option 1. Paul Weiss also prefers Option 1, because the result would parallel $v authority records. The Research Libraries Group does not support Option 1, because there will not be enough tags if later separate tags are requested for function, curriculum, occupation, etc. Priscilla suggested the possible lack of enough free tags be addressed separately. Apparently, Toni Peterson from AAT questions whether there would need to be duplicate records in the authority file if Option 1 were used. AAT has a policy of using the same heading if already used by another thesauri. This may result in duplicate validation if the same term is used for both topic and form. Paul Weiss said that the 008 field would tell people. Sherman Clarke pointed out that AAT does not try to control use, just the vocabulary form. Arlene Taylor pointed out that LCSH has some dual terms, and different X-refs may be necessary; this fact supports the need for two authority records. On page 9, the $v is used as a place for the Record Control Number. Given that 95-2 was passed earlier, Priscilla said that it is clear that another subfield is needed. It was agreed to move the issue to the proposal stage. The proposal should use Option 1, with development of 008/18 in relation to it. Pros and cons should be discussed. It was also agreed that the occupation and function terms should NOT be considered as part of this proposal. To everyone's surprise and delight, the meeting ended 45 minutes early! ********************************************************************* February 6, 1995 ********************************************************************* ANNOUNCEMENTS: Larry Woods is chairing the MARBI Unicode Subcommittee; other members include Joan Aliprand (RLG), Randall Barry (LC), John Espley (VTLS), Robyn Greenlund (MicroLif), Gary Smith (OCLC), and Paul Weiss (NLM). Larry reviewed the principles under consideration: round trip mapping of characters, preserving the identity of USMARC characters, transliteration tables should remain unchanged, want to continue to follow the practice of decomposed characters, and mapping punctuation and numbers to a UCS repertoire. The Subcommittee has completed mapping for Basic and Extended Latin characters sets, and will next tackle Hebrew, Arabic, etc. Larry wants to set up a schedule for completing the different parts of the project, and will post a written report on the USMARC list in the near future. Sally McCallum brought up the issue of SGML mark-up language and its relationship to USMARC records: 1) It might be possible to append USMARC records to SGML documents in some consistent manner. 2) If electronic books are coded in SGML in a consistent fashion in the future, could a cataloging record be generated automatically? LC is looking for funding to pull together experts of both USMARC and SGML for discussion. Priscilla Caplan reported that OCLC is hosting an invitational meeting on "metadata" in early March. Movers and shakers of the Internet community, the Text Editing Initiative (TEI), and the visual imaging community are invited in addition to library folk. Will try to reach a consensus on how to describe and prioritize electronic resources, and will deal with the affect on USMARC, SGML, TEI, etc. A full report will be prepared for MARBI for discussion at the June meeting. ----------------------------------------------------------------- PROPOSAL 95-4 : "Merger of the 27X fields in the USMARC Format for Community Information" Rebecca Guenther introduced Proposal 95-4. Recently Field 270 was added to the bibliographic format (see Proposal 94-9). Would now like to align the two formats by using Field 270 only in the Community Information Format and making Fields 271 and 275 obsolete. Would also add a $r and $z to Field 270 to be equivalent to the field in the bibliographic format. Lee Ireland, representing PLA/CIS, pointed out that the needs between the bibliographic and community information communities needs are different. Asked MARBI to examine the concept of complete alignment to the bibliographic format. One example is the following: Primary address (a recreation dept) 10-25 Secondary addresses (parks, branches, auxiliary service points, different mailing addresses, etc.) Lee is not at all sure if all address needs will be met if the proposed change occurs. It must be possible to easily update addresses. It must be possible to easily organize like addresses. This gets complicated. One must also take into account what is already present in the 1xx/245 fields. Another member of the audience spoke up with display concerns. Won't it be more programming if the type of address must be checked for display purposes? The type of address is buried in the $i. If separate fields are not possible, another indicator would be helpful. Priscilla Caplan asked if it would help if the indicators were re-defined from primary/secondary to agency, program, mailing address, etc. John Espley replied that both primary/secondary address coding is needed, as well as type of address like program, agency, mailing. The community information people would like a more flexible approach to address types. Robyn Greenlund suggested that the mailing address be the default. Paul Weiss had a different suggestion: let the first indicator be the agency or program, and the second indicator show if a mailing address or not. Diane Hillmann pointed out that home/office addresses may be needed, because a person's name may be in the 100. She suggested that specifying at the indicator level may be too restrictive. Bill Jones asked if the $i should come first, to facilitate its use (the 246 is a precedent for this). Larry Woods did not agree. And, John Attig suggested that indicators be used wherever the most computer processing will be done. Marti Scheel pointed out that the universe of options has not been covered in the first indicator code. Priscilla suggested adding a "blank" to mean "no info available." The PLA group would like to talk with more people, to come up with a better proposal. Perhaps the hours should be in a $v, not $r. Definitely want to consider different indicator schemes, and probably include "blank indicators" to provide more flexibility. John Attig added that it is nice to be in parallel with the bib format, but he does not feel that it is always a practical idea. John Espley added that users would very much like to have input guidelines, to help them use the CI format more easily. Priscilla asked a question about the $4 relator code. What would this code show a relationship to? Rebecca responded that the $4 would contain a code for the type of address, whereas the $i would show the free-text form of the type of address. Rebecca further added that relator codes can be added easily to USMARC. It was agreed to rework the proposal, taking into account the PLA/CIS group input. The distinction between primary and secondary addresses needs to be maintained. Examples would be very helpful in the next proposal (probably in missing pages from Attachment A). LC NDMSO will rework as a proposal with options after PLA advises. ----------------------------------------------------------------- PROPOSAL 95-7 : "Change of tag 301 to 307 in the USMARC Format for Community Information." Sally McCallum introduced the proposal and characterized it as being somewhat straightforward. As a result of changes to the bibliographic format to accommodate online systems and services, a field was added to the bib format to record hours. Field 301 was initially suggested, to be consistent with the Community Information format. But, this was not possible since 301 had already been used. Since obsolete tags cannot be reused with a different definition, 307 was approved in the bib format. It is now proposed to change the provisional 301 field in the Community Information Format to tag 307. Sally pointed out that some users want the two formats to mesh together; this will cut down on systems work, allow a system to load both types of records in the same file or database if desired; or support cross-file searching if separate files are preferred. Robyn Greenlund pointed out that not all format alignments will be worth people's time. John Espley asked if the proposal had been discussed as of yet by the Community Information MARC Format Discussion Group. Gary Strawn made a motion to approve the proposal as is. His motion was seconded. There were six votes in favor and no votes against. ----------------------------------------------------------------- PROPOSAL 95-6 : "Linking Code for Reproduction Information in the USMARC Bibliographic Format" Sally McCallum introduced this proposal by reminding attendees that the $8 linking technique was approved at an earlier meeting, with the caveat that specific linking codes will be approved separately. If this proposal is approved, it will be the first instance of a linking code, in this case for reproductions. The proposal assumes the cloning technique, where the reproduction record is a clone of the original. After cloning, a library would add the reproduction descriptive information to the second record, marked by the $8 linking technique ($8 followed by the link number followed by a backwards slash followed by the "r" link code). John Attig asked if the expectation is that everyone will code reproductions in this way, if the "r" code is approved. Or, is the expectation that the linking would be optional? Young-Hee Queinnec believes that the approach should be optional. Paul Weiss agrees; he thinks it should be left up to local library policy. Sally McCallum brought up the problem of existing records. There are many records with 533 and 007 now for reproductions. The grandfathering problem must be addressed, ideally where the new approach will be as consistent as possible with existing records. John Attig is looking for a tool where libraries can maintain separate descriptions. He wondered if there has been adequate exploration of all possibilities. For instance, some people use the 843 field in holdings records now. Josephine Crawford explained that this indeed is the approach used at the University of Wisconsin, Madison. The public services staff much prefer a single, master bibliographic record based upon the original with descriptive information for both original and reproduction. The reproduction information is then put partly into the bibliographic record (series entry, for instance) and partly into the holdings record (007 and 843 fields usually). She believes the linking technique will improve machine control over the master record, but questioned the assumption that records would be cloned. This seems unnecessary in a local system, if the local preference is for a master record. The linking technique will allow input/output programs to merge or split apart as needed for record communication. Paul Weiss stated that NLM plans to follow the single record approach. The earlier proposal (94-15) had a single record example. Apparently, there were different interpretations because he was surprised to see the assumption of cloning. John Attig pointed out that there must be a holdings record attached if a master bib record approach is used, to show both the original and microform locations and call numbers. Jo Crawford indicated that this is done at the Univ of Wisconsin, Madison. Both Diane Hillmann and Paul Weiss had posted examples to the USMARC list. It was agreed that more examples would be helpful. John Espley spoke up with a concern. Are we backing off from the multiple versions decision? He reminded the audience that many systems do not support the holdings format yet. Robin Wendler reminded participants that separate records had been agreed upon, after much back and forth. Larry Woods wondered if we are stumbling over what we communicate versus what we store in a local system. Catherine Gerhart stated that it is not clear at all, but she prefers local system flexibility. Rich Greene pointed out that record exchange is multi-directional. What might be good for a local system may not work for the utility. Priscilla added that research libraries want to do more with less dollars. She believes that massive quantities of records will be accepted in the future with no machine or hand editing. John Attig said let's not belabor the two records versus single record issue. Sally McCallum said that the proposal is trying to accommodate the CC:DA microform guidelines and the large numbers of 533 and 786 fields in existing records. Others asked how to handle the 533 and 007 fields, for multiple reproductions plus the original. By repeating them? Already a messy situation. Another problem is the GMD in the 245 field. Priscilla Caplan suggested that the $h GMD be stripped from the record. John Attig reported that the CC:DA guidelines has the original GMD in the 245 field, so the examples are in error then. Bruce Johnson reported that the guidelines were finalized nine months ago and are just now in print. He found that original and reproduction information has been mixed in the same field. CC:DA would like all reproduction information to be pulled together; a 533 field is OK but other fields are needed too. Would it be possible to have multiple 245 fields? Robyn Greenlund suggested putting the 014 field into the bib format. The microform record would only have micro information in it, and the 014 would contain a matching control number. Systems would merge multiple records for display. John Attig believes that MARBI decided long ago to not go with the 014 approach. Sally wondered if the 786 link to the other record isn't the same technique. Others agreed it is. Should the 533 field be required? Should all subfields be dropped except for the $a, or keep the same subfields? Should other subfields be added? Sherman Clarke pointed out that the CC:DA quidelines show that more than 26 data elements are needed, so more than 26 subfields are needed. Others agreed that there might be display problems if all reproduction information is kept in the 533. Also, most systems don't index note fields or don't index them in the same way. To try to wrap up the discussion, Priscilla Caplan decided to take a number of straw votes, asking everyone in the room to participate. 1) Does anyone object to separate records for communicating a reproduction and an original? Most voted no, a few voted yes. 2) Is it a good idea to have a way of marking version information on records? Yes, a unanimous vote. 3) Is the $8 linking technique a good approach? Yes, unanimous. 4) Should the $8 be required for all reproduction fields, including the 533, but making an exception for the 007? Most say yes, some say no. 5) Just to be clear, should the $8 be required for the 533 and 776, or assumed as occurs now? Most say yes, two say no. Diane Hillmann pointed out that the environment is growing in regards to reproductions. To be relevant to our users, we must deal with this issue. Sherman Clarke emphasized that we must find a way to link different reproduction descriptions within a record. Catherine asked if $8 is a link, a marker, or both? How will the $8 work if only a 533 field is present in a record, or if the reproduction data is represented by a set of fields, or if there are multiple reproduction descriptions in the record? Priscilla replied that all three situations will occur. To her mind: -- No need for $8 if only a single 533 field in a record and it contains the entire reproduction description. -- Supply $8 if there is a 533 plus other fields for a single reproduction description. -- Supply $8 if there are multiple reproduction descriptions in the same record, using the link number as appropriate. Sally envisions $8 links of different types in the same bibliographic record. Must have a link number and a link type code. In terms of the 533 subfield issue, Priscilla doesn't think we can address this at this time. John Espley asked if a back slash is required, because VTLS uses this to indicate a subfield. Yes, it is required, as approved at last MARBI meeting. It was agreed that the proposal will be modified for the June meeting in Chicago, reflecting today's discussion. ----------------------------------------------------------------- DISCUSSION PAPER #81 : "Form of Music Code in Bibliographic Records" The National Library of Canada has requested new codes. Some LC staffers think that these codes are redundant with 6xx form subdivisions and the 655 field. The music cataloging community is not unanimous on the issues of musical codes versus form of composition codes. Some music catalogers want to add new codes, others want to do away with them altogether. Right now the codes go into the 008/18-19, and field 047 when more than one form of composition is involved. The list of codes is incomplete at this time; LCSH has established new musical forms, but the USMARC list of codes has not been changed. If codes are kept, do we want to have a one-to-one correspondence with LCSH? John Attig believes that some people see potential value for the codes. Karen Little knows that a few local systems use the codes for access. She suggested putting a notice out on several lists. Many think codes should be obsolete, but a handful of systems do make use of them. A Music Library Association working group has come up with a list of 50 potential codes, but they raise many questions. For example, why not "lieder" in French and Italian? The terms are not consistent unfortunately. The music thesaurus project may provide a better textual term list in the future. Young-Hee Quiennec recently forwarded input from the Canadian music group to LC. Costs should be considered. John Attig pointed out that the balance point between codes and text is changing; there is a cost to put in codes and it is getting easier for systems to deal with textual strings. Priscilla thinks that it might be best to make these codes obsolete, but it is important to hear from the bibliographic community first. Sally really encouraged everyone to discuss and send input to her office, but especially the Music Library Association and the Canadian MARC Committee. There was a general consensus that further investigation is needed before making these codes obsolete. If the codes are not made obsolete, then better maintenance is required. It was also agreed that codes need not be added by proposal; they can be added to one of the existing code lists more quickly.