Josephine Crawford ALCTS University of Minnesota James Crooks RUSA University of California, Irvine Elaine Henjum LITA Florida Cntr. for Library Automation Diane Hillman LITA Cornell University Carol Penka RUSA University of Illinois Jacquie Riley RUSA University of Cincinnati Frank Sadowski ALCTS University of Rochester Paul Weiss ALCTS University of New Mexico Robin Wendler LITA Harvard UniversityMARBI Interns:
Annemarie Erickson RUSA Chicago Library System Anne Gilliland (recorder) ALCTS OhioLINK Chris Mueller LITA University of New MexicoRepresentatives and Liaisons:
Joe Altimus RLG Research Libraries Group Karen Anspach AVIAC Eos, Intl. John Attig OLAC Pennsylvania State University Sherman Clarke VRA New York University Betsy Cowart WLN WLN, Inc. Donna Cranmer ALCTS Media Resources Cmte. Siouxland Libraries Bonnie Dede ALCTS CCS SAC University of Michigan Catherine Gerhart ALCTS CCS CC:DA University of Washington Kathy Glennan MLA University of Southern California David Goldberg NAL National Agricultural Library Rich Greene OCLC OCLC, Inc. Rebecca Guenther LC Library of Congress Brian Holt BL British Library Michael Johnson MicroLIF Follett Co. Maureen Killeen A-G A-G Canada Ltd. Rhonda Lawrence AALL UCLA Law Sally McCallum LC Library of Congress Susan Moore MAGERT University of Northern Iowa Elizabeth O'Keefe ARLIS/NA Pierpont Morgan Library Louise Sevold CIS Cuyahoga County Public Library Marti Scheel NLM National Library of Medicine Margaret Stewart NLC National Library of Canada Rutherford Witthus SAA University of ConnecticutOther Attendees:
Jim Agenbroad Library of Congress Everett Allgood New York University Randall Barry Library of Congress Jack Cain A-G Canada Mehmer Celik ELIAS Vinod Chachra VTLS Karen Coyle California Digital Library Ellen Crosby Indiana Historical Society Robin Dale RLG Harriet DeCeunynd Rutgers Paula DeStefano NYU John Espley VTLS, Inc. Bernard Eversberg Universität Braunschweig, Germany Charles Gordon University of South Florida Jane Grawemeyer SIRSI Stephen Hearn University of Minnesota Jean-Frederic Jauslin Swiss National Library Bruce Chr. Johnson Library of Congress Jane D. Johnson UCLA Film and Television Archive William Jones New York University Shirley Kieran Best Seller Ken King UMI Deborah Leslie Yale University Elizabeth Mangan Library of Congress Christina Meyer University of Minnesota John Riemer University of Georgia Marie Robertson Book Wholesalers Inc. William Russell GEAC Donnell Ruthenberg Data Research Associates Mary Schneider Catalog Card Company Jacque-Lynne Schulman National Library of Medicine Ann Sitkin Harvard Law Library Gary Smith OCLC Karen Smith-Yoshimura RLG Steve Squires UNC-Chapel Hill Daniel Starr MOMA Barbara Story Library of Congress Gary Strawn Northwestern University Bob Thomas WLN Grace Thomas University of California, Santa Barbara Russ Thompson BRODART David Walker Follett Library Resources Bob Warwick Rutgers University David Williamson Library of Congress Bob Wolven Columbia University Ruth Wuest Endeavor Joe Zeeman CGI Group
AALL - American Association of Law Libraries
ALCTS - Association of Library Collections and Technical Services
ARLIS/NA - Art Libraries Society of North America
BL - British Library
CC:DA - Committee on Cataloging: Description and Access (of ALCTS CCS)
CIS - Community Information Section (of PLA)
CCS - Cataloging and Classification Section (of ALCTS)
LC - Library of Congress
LITA - Library and Information Technology Association
MAGERT - Map & Geography Roundtable
MLA - Music Library Association
NAL - National Agricultural Library
ND/MSO - Network Development and MARC Standards Office (of LC)
NLC - National Library of Canada
NLM - National Library of Medicine
OLAC - Online Audiovisual Catalogers, Inc.
PLA - Public Libraries Association
RUSA - Reference and User Services Association
SAA - Society of American Archivists
SAC - Subject Analysis Committee (of ALCTS CCS)
VRA - Visual Resources Association
Saturday, June 27, 1998
Jacquie Riley, MARBI chair, opened the meeting with introductions. The list of committee members was passed around for corrections. This list will be made available on the MARBI web site.
Proposal 98-8: Coding of Exceptions to Regularity Patterns in Field 853-855 in the USMARC Holdings Format
Rebecca Guenther introduced the proposal. Since 1992, when changes were made to $y in fields 853-855, there have been many occasions when there have been further questions and discussions about coding regularity patterns. This has been a particular problem because $y is non-repeatable, so it is only possible to give one regularity pattern or one exception. Rebecca pointed out that the example in paragraph 2.6 has an error; "1201" should be dropped from the first $y.
Proposal 98-8 gives two options. Option 1 is to define a new repeatable subfield, $r, for regularity exceptions. Option 2 is to make $y repeatable. In the discussion that followed, initially there was no clear consensus on Option 1 or 2, with some preferring a combination of both options. Others pointed out that the idea of making the $y repeatable had been rejected in the past because it could cause ambiguity. Paul Weiss pointed out that the last sentence in paragraph 2.6 was not correct, because at that point it is not possible to tell whether the alternate pattern is in addition to, or as a substitute for, the first regularity pattern.
Rebecca Guenther commented that if there is a point where regularity exceptions become so frequent, then the pattern needs to be changed. There was also discussion about whether code d can be used for numbers of days as well as names for days of the week. LC may clean up the appendix to the holdings format because of a number of errors and ambiguities in the examples.
Paul Weiss moved that Proposal 98-8 be adopted, using Option 2, noting that sentiment was pointing toward that solution. Diane Hillman seconded the motion. The motion passed, with 9 voting for, none against.
Proposal 98-9: Charges to the USMARC Classification Format for Number Building
Rebecca Guenther introduced the proposal, which has come out of activities to convert the Library of Congress Classification schedules into machine-readable form. There is nothing in the Classification Format that indicates what rule to use in the 7XX Number Building fields. Adding this information to a standard MARC field would have no known impact on other classification schemes, and could be useful for users exchanging records.
There was some discussion about whether both $a (rule number) and $i (explanatory text) was necessary. The explanatory text is not strictly necessary, but it is useful for classifiers and for those checking the work of calculator functions. This change has already been implemented at LC.
Paul Weiss moved that the proposal be approved, with Robin Wendler seconding the motion. Proposal 98-9 passed, with 9 voting in favor, and none voting against.
Proposal 98-10: Definition of Subfield $0 for Record Control Number in the 7XX Fields in the USMARC Classification and Community Information Formats
Rebecca Guenther introduced the proposal. As part of the alignment of USMARC and CAN/MARC formats, $0 for record control number was defined for the 5XX and 7XX fields in the USMARC Authority Format. After analysis, it was determined that it would be best if the subfield for authority record control number in the 7XX fields of the classification and community information formats were also $0.
After some discussion, Paul Weiss moved that the proposal be accepted with the change that $w be deleted, instead of being made obsolete. Diane Hillmann seconded the motion, which passed unanimously.
Proposal 98-7: Recording Incorrect Dates in Field 008/06-14 in the USMARC Bibliographic Format
Rebecca Guenther introduced this proposal, which arose from Discussion Paper 106 (January 1998). It proposes the implementation of a new type of date code in the bibliographic format to provide access to incorrect publication dates that appear on items. These codes would be used in field 008/06-14 and field 046.
John Attig said that this proposal meets the needs of the community that first raised the problem of recording incorrect dates. Discussion then turned to an amendment from RLG, suggesting that the 008 be left as is, and that the 046 be the only place where incorrect dates are recorded. In this scenario, the code "x" would be used to indicate the presence of incorrect dates, and $f and $g would be added for recording the incorrect dates. Many saw this solution as problematic, since no one knew of a system that uses the 046 for processing. John Attig pointed out that RLG's amendment was workable, but only if systems' functionality are improved. There was also discussion of how systems also do not fully use the 008 for processing. For example, if there are two dates in the 008, some systems do not process them as a range (when appropriate). There was also some concern that it is difficult to train staff to look for this information in two places. It was pointed out that it is a mistake to tie format changes to the way systems currently work. On the other hand, OCLC does not index the 046, and has no plans to do so soon. There was also some discussion about whether, in RLG's scenario, it would be necessary to use 008/06 code "x," since the presence or absence of $f and $g would signal whether incorrect dates were recorded. It was agreed that it was not necessary to use 008/06 code "x."
Jacquie called for a straw poll on RLG's proposed amendment using only field 046 for incorrect dates. The majority voted for RLG's solution without the use of code "x" in field 008. After more discussion, Diane Hillman moved to accept RLG's proposal, except with the use of letters other than "f" and "g" for the subfields. Paul Weiss seconded the motion, which passed unanimously. RLG will follow up with a proposal for date information from non-Gregorian calendars. ND/MSO will review the 046 field to determine whether there is a need for new subfields $f and $g or whether to use existing subfields $c and $e and will report back to MARBI.
Proposal No. 98-15: Obsolete Fields in the USMARC Bibliographic Format
Sally McCallum introduced this proposal, which recommended making the following data elements obsolete: 261, 400, 410, 411, 260 $d, X11 $q, and 262. This would complete the harmonization of the USMARC format with CAN/MARC. Rich Greene (OCLC) pointed out that there is quite a bit of retrospective conversion activity, and that some of the fields being considered in this proposal are still used as a part of retrospective conversion projects. Since OCLC either lists a field or element as valid or invalid, the distinction of being obsolete is not completely useful in that environment. He stated that OCLC staff were split on the desirability of the proposal. Others commented that the 4XX fields in question allow for machine flipping to 8XX fields. Kathy Glennan (MLA) pointed out that the 262 does not map neatly to another field. John Attig observed that these fields were left alone during format integration because of retrospective conversion projects. He said the AV Online Cataloger group that he represents could accept the change. Margaret Stewart from NLC pointed out that CAN/MARC has never included the 262 field.
After further discussion, Paul Weiss moved that the proposal be rejected. Diane Hillman seconded the motion. The motion to reject the proposal passed with eight voting in favor and one voting against.
Proposal 98-17: Reading Program Information in the USMARC Bibliographic Format
Rebecca Guenther made some preliminary comments about this proposal, and then asked Michael Johnson to finish introducing it. This proposal arises from Discussion Paper 105 (December 1997) that was considered Midwinter 1998. At that time, MARBI indicated that they would favor the use of a single field approach using the 526 field to record reading program information.
Joe Altimus from RLG suggested that a display constant be added to the indicator so that the 526 could be more broadly defined for all sorts of curriculum-based programs. There was also some discussion about using the 521 instead of the 526 for this information. Others replied that it was preferable to keep this information in a separate field so that it could be filtered out of records if desired. There were also concerns about the ephemeral nature of reading programs that are not structured commercial programs. The use of $5 might be helpful in those cases.
Michael Johnson answered the questions in section 4:
4.1. Does the reading program need to represent a physical object, or can it be like a program put together with a theme (e.g., Oprah's book club)? No
4.2. Since the answer to 4.1 is No, this question is not applicable.
4.3. Should subfield $5 be defined for Institution to which field applies? Yes
After further discussion, Diane Hillman moved to accept the proposal with the following modifications:
Carol Penka seconded the motion. The motion was approved 8-0.
Sally McCallum announced that the new print version of the MARC Concise Format is now available, as is a new edition of the USMARC Code List for Geographic Areas (GAC) list. LC is working on new editions of the Bibliographic and Authority formats with NLC, as well as an update to the Holdings Format. Next year they will issue new editions of the Classification, Community Information, and Holdings formats. LC is adding the full Country, Language, and GAC code lists to the web. They will still maintain the ASCII versions on the Web for institutions that use them for validation. All proposals and discussion papers are now distributed via the web. LC is aware that some people are having trouble printing discussion papers, and they will be changing formats to eliminate this problem. LC is in the midst of changing the browsing and searching mechanisms for the listserv archives.
There were questions about proposals that have been approved, but which have not yet been implemented. This causes confusion if the utilities implement them before LC. Sally replied that LC handles these as special announcements.
There was an announcement of the joint CC:DA/MARBI meeting on Monday. Discussion topics will include cataloging metadata, rules issues, the definition of seriality in the cataloging rules, and carrier vs. content issues. The CC:DA Metadata Task Force, which includes MARBI members, will meet tomorrow. Rebecca Guenther announced a forum to be held by the MARC Formats Interest Group on "MARC Formats: A World View." Jacquie Riley had received a letter from RUSA asking for MARBI's support for work on standardizing search commands. The committee agreed that this topic was outside MARBI's scope. Jacquie also announced that Bruce Chr. Johnson from the Library of Congress will be the next MARBI chair. She expressed her appreciation and gratitude to the committee for their support during her tenure.
Sunday, June 28, 1998
Proposal 98-12: Additional Indicator Value in Field 355 (Security Classification Control) of the USMARC Bibliographic Format
Betsy Mangan from LC's Geography and Maps Division introduced the proposal, which is a result of work with the U.S. National Imagery and Mapping Agency as they move to the MARC format. This agency has identified a need to add an indicator to field 355, which is used for security classification control. They need to be able to show whether a record is classified information, whether the material described is classified, or both. This will allow them to tell whether the record can or cannot be communicated to other agencies.
Paul Weiss moved that the proposal be accepted; Robin Wendler seconded the motion. The motion to accept the proposal passed 7 - 0.
Proposal 98-14: Additional Code List for Field 052 of the USMARC Bibliographic Format
Betsy Mangan introduced this proposal, which comes out of work with the U.S. National Imagery and Mapping Agency (NIMA). It suggests that field 052 (Geographic Classification Code) be broadened to allow for recording codes for entities other than those coded in the Library of Congress Classification schedules. This arises from NIMA's need to use a code list from the Federal Information Processing Standards Publication 10-4 (FIPS 10-4).
The proposal gives the choice of two options. The first is to define the first indicator value for the source of code with # for LC Classification and 7 for source specified in $2. The second is to leave the first indicator value undefined and always use $2 for the source of the code. Both options propose changing the name and description of the field and changing the subfields to allow for codes from other sources.
There was discussion about whether the 043 field should be used for this information instead of the 052. It was noted that the 043 field is bound by the rules of the GAC. Some commented that option 2 is more flexible, with others questioning how much flexibility is actually needed for coding this information. NLC supports Option 1 because some information is specified at the beginning of the field. It was suggested that there be a separate indicator value for Dept. of Defense. Although defining separate values could result in running out of them, a call for other geographic classification schemes on the Maps-L list did not identify others needed.
Paul Weiss moved that the proposal be passed with the following modification to Option 1: Indicator 1 - value # - LC Classification; value 0 - U.S. Dept. of Defense Classification. Carol Penka seconded the motion. The motion to approve the proposal as amended passed 7 - 1.
Proposal 98-13: Defining Field 856 in the USMARC Authorities Format
Rebecca Guenther introduced the proposal. The 856 has been defined in the Bibliographic, Classification, and Community Information formats. Its purpose in the authority record would be to add supplementary biographical and historical information. The committee's reaction was mixed when this idea was submitted as Discussion Paper 107 (January 1998). Rebecca also introduced John Riemer, who had originally brought forward this idea.
Robin Wendler pointed out that the definition of $3 is problematic because a URL often links to a finding aid. Rebecca replied that the use of $3 has not been consistent, and may need a comprehensible note. John Riemer pointed out that one can use $3 and $z to make the relationships more obvious. There was some discussion of linking to "non-official" web sites. For example, section 2.5 of the proposal has an example of a link to a photograph of Bertrand Russell (as opposed to an "official" Bertrand Russell web site). Is this the purpose of 856 fields in authority records? Will there be limits to this sort of linking? Diane Hillmann noted that not specifying the appropriate use of URLs in the Authority format would be a mistake and could add to libraries' maintenance burdens. Robin Wendler added that the entities that maintain thesauri might appropriately make these specifications. Others raised the issue of problems with socially sensitive topical headings. Marti Scheel pointed out that NLM would like to have MARBI define the 856 for the Authority format, because that will drive vendor development in this area.
Sally McCallum said that LC would employ examples in its documentation that would take a conservative approach to the use of the 856 in authority records. Bonnie Dede said that SAC had discussed this issue and had raised many of the same concerns that MARBI members have. Sherman Clarke agreed that VRA members were concerned about maintenance of the records, but thought that the concept would be important for cataloging visual images. Others pointed out that the use of 856 fields in authority records increases their usefulness. There was discussion about when a 856 field should be employed in an authority record instead of a bibliographic record. There was also some sentiment in favor of using 856 fields in authority records instead of in community information format records, since the CIF format has not been well integrated into catalogs. Committee members felt that additional discussion on the use of subfield $3 was needed because the application of the subfield in authority records is ambiguous. This discussion will be continued on the USMARC listserv. Questions arose about how the second indicator would be used; consensus was that it would always be blank.
Frank Sadowski moved that the proposal be approved. Elaine Henjum seconded the motion. The motion to approve Proposal 98-13 passed with 7 voting in favor, 1 voting against, and 1 abstention.
Discussion Paper No. 110: Enhancement of Computer File 007 in the USMARC Bibliographic/Holdings Formats
Rebecca Guenther introduced the discussion paper. She introduced Robin Dale, representing an RLG preservation issues working group that initiated this discussion. The working group's concern was with recording information about digitally reformatted items. The 856 field is not adequate for identifying them because of the need to show the level of preservation. This in turn demonstrates institutional commitment to continue to provide access. Marti Scheel (NLM) asked whether RLG had talked to the National Archives and Records Administration and other similar groups. Robin Dale said "no," although the discussion had been posted on web sites and listservs within the preservation community.
Discussion then focused on the numbered questions in the discussion paper.
Question 1 (It would be difficult to enforce that they be mandatory for a specific type of computer file. Would it be preferable to propose that they be highly recommended for preservation computer files? ): It was agreed that the enhanced 007 could not become mandatory for certain kinds of computer files.
Question 2 (Could it be considered to establish a new 007 for preservation computer files instead of using the established one? In this case the elements that exist in the CF 007 could be repeated and these additional bytes added. This might be a cleaner method so that the original 007 for computer files is not affected. How have users been served by the precedent mentioned above, i.e. adding the bytes for archival film elements to the 007 for motion pictures?): RLG is willing to consider a new 007 instead of using an established one.
Question 3 (Varies has been defined as the item will change over time. If it's coded as "v" then is it assumed that one must go back to change the information in the other bytes when the file changes (e.g. dimensions, image bit depth)? Doesn't saying it varies means that the physical details should not be recorded? Or would the cataloger keep adding new 007s to cover physical details when the file changes? What is the purpose of the "v" if you either do not give physical details or go back and add additional 007 fields? Why doesn't "other" suffice? If it is found that varies is indeed needed, perhaps a better term would be "dynamic." In addition if a separate 007 were used for preservation computer files, then defining "v" wouldn't confuse what's already there (although maybe it could be assumed and be no longer needed).): Yes, it is necessary to change information when the file changes. "Other" is not used because it denotes some unnamed fixed format. "Dynamic" is a better term than "varies."
Question 4 (This byte does not seem to cover audio/visual or OCR but only scanned items. Should it be made clear that non-scanned is not included here? For instance the Prints and Photographs Division at LC considers the print of the photographer the original, and if they do not have that they consider the negative the original. The definition does not define what is meant by "original." What if it's scanned from a photocopy? And how to code it for a non-2 dimensional scanned image, e.g. video?): "Antecedent" and "source" are not synonymous; there is some disagreement within RLG about what is being coded. This language and the intention behind it needs to be clarified.
Question 6 (Can this always be limited to two bytes?): No, 007/08-09 (Image bit depth) could go to three bytes.
Question 7 (Does this indicate that quality assurance has been done? Quality assurance does not have to be done through targets, and monitors can be adjusted according to a known target. In addition this may not cover all forms of practice in the future.): No, this does not indicate that quality assurance has been done.
Question 8 (It might be preferable to change the codes so that they are not mneumonic, since they can't be (lossless and lossy). Many of the codes in the computer files 007 are just in order alphabetically, so these could be changed to a, b, c, etc. Must the distinction be made between not applicable and uncompressed? Can't someone always come up with a way to compress a file? Can not applicable be deleted? What if there is a mixture of compressions? Should a value be added for mixed or combination? Would a searchable text file be coded as uncompressed? Levels of compression can be controlled so that a lossy technique is used, but the result is essentially a lossless compression. How would this situation be coded?): It is not clear that it is possible to code a mixture of compression techniques. RLG agrees to making codes non-mnemonic. Not applicable could be omitted, and it might be considered that "mixed" be added.
Jacquie Riley called for a straw poll on the use of a separate vs. an expanded 007. No one voted in favor of a separate 007. More voted in favor of an expanded 007, but many also abstained. Jacquie ended discussion due to a lack of time. LC ND/MSO will post Question 9 and the remaining general questions to the USMARC listserv. Based on the results of this discussion, ND/MSO will work with RLG to develop a proposal for 1999 ALA Midwinter in Philadelphia.
Proposal 98-18: Unicode Identification and Encoding in USMARC Records
Gary Smith introduced this proposal on behalf of the Unicode Encoding and Recognition Technical Issues Task Force. It addresses technical issues remaining unresolved from Proposal 97-10 (June 1997), including encoding of Unicode characters in USMARC communications records and a method for identifying the use of Unicode in USMARC Communications records.
There are two options for identifying records encoded in Unicode. Option 1 is to use Leader/09 for this purpose; Option 2 is to use Leader/23. Sally McCallum spoke in favor of using Leader/09 to remain in compliance with ISO 2709. She also made the point that this byte will not be optional, but the encoding in Unicode will be optional. Implementation is up to community and exchange partners.
The Task Force did not make a decision on encoding double diacritics (Issue 4) because of the variation of practice in existing records. This has a bearing on problems raised in Proposal 98-16. Committee members commented that position counts and lengths (Issue 5) should be carefully documented. There was a suggestion to re-word 3.1.h to note that this is not a requirement of MARBI. When LC does this, they will propose nothing that is contrary to Unicode.
Diane Hillmann moved to approve Option 1 of Proposal 98-18. Paul Weiss seconded the motion. The motion to approve Proposal 98-18 passed unanimously.
The Task Force was asked to continue its work so that the questions and issues raised in Appendix A are discussed and resolved. Items 2 and 3 in Appendix A will be given highest priority (and it was noted that item 6 was actually no longer needed as it is treated in 98-18). RLG will study the issues raised in Item 5 of Appendix A.
Proposal No. 98-11: Changes to the USMARC Holdings/Bibliographic Formats Resulting from the New Holdings Standard (Z39.71)
Rebecca Guenther introduced the proposal, which is an effort to keep USMARC in line with Z39.71. Proposed changes are a new 007 for kit, a new 007 for music (exclusive of sound recording), the addition of value "u" to 007/01 for all 007 fields except computer files, a change of value # to u in 007/01 of Remote sensing image, and a change of name for the Map 007 to "Cartographic Material."
There was discussion about the term "music (exclusive of sound recording)." MLA will discuss an alternate wording; musical notation was suggested. "Cartographic material" is not the appropriate change from "Map" because globes and atlases are also cartographic materials, and they have their own 007s.
Frank Sadowski moved that MARBI approve Proposal no. 98-11, with the exclusion of the change for maps. MLA was asked to consider the appropriate name of this 007 for music materials. Paul Weiss seconded the motion, which passed unanimously.
East Asian Character Set Task Force
John Espley, Task Force Chair, reported on the difficulty mapping this character set to UNICODE. There are problems with round-trip mapping, and the Task Force anticipates using the private use space to clarify some of the problems. The Task Force plans to bring a proposal to the MARBI meetings at 1999 ALA Midwinter in Philadelphia.
John Attig reported on Saturday's meeting. German libraries have made a proposal that could have an impact on USMARC. He will post the URL for the relevant document on the USMARC listserv. CC:DA is also having ongoing discussions about seriality, which may ultimately have a bearing on the MARC format.
Monday, June 29
Joe Altimus passed out a handout that shows some examples of the issues presented in DP 110 in context. There had not been enough time to discuss them the day before.
Proposal 98-16: Nonfiling characters in all formats
Sally McCallum introduced the proposal, which follows Discussion Paper 102 (June 1997) from the 1997 ALA Annual meeting. The increased use of the 246 field has made the need for a nonfiling indicator more urgent. This is also a recurring problem in names and in fields that use $t. Paragraph 1.3 of the proposal summarizes the requirements for any changes to current practice.
This proposal takes the position that a control character should be used to show the nonfiling requirements, because this is what is used most often throughout the world. It is anticipated that systems would adopt some eye-readable convention for showing them, but there are no obvious keyboard equivalents. Currently, other control characters, such as the end of file mark, the delimiter, and the end of record mark, are represented by various visible symbols.
Karen Coyle reiterated some of the points she had made in her June 18th message to the USMARC listserv, stressing the difference between filing and sorting on one hand and retrieval on the other. This spurred a discussion about how systems do or do not handle articles when sorting. To further complicate matters, in some languages articles come after the words they modify. Bernard Eversburg (Universität Braunschweig) commented that many years' experience in Germany leads him to believe that it is best to use one control character, which surrounds whatever is supposed to be ignored for filing or sorting. Vinod Chachra (VTLS) commented that he would prefer to implement two different characters in case there are problems discerning where the non-filing text begins and ends. Gary Smith (OCLC) concurred. Representatives from OCLC and RLG said that this would be a major change to implement, but that they could see many advantages in doing so.
Discussion then turned to the impact of this change. Rich Greene (OCLC) pointed out that utilities will need considerable lead-time, and that decision will be needed about converting records. Randall Barry (LC) reported that the Unicode consortium, ISO 10646, is interested in what MARBI decides about this issue, but will not accept the same solution immediately. Nevertheless, the use of two different control characters will probably be more compatible with Unicode. This led to more discussion of using one character or two, with some saying that logically it should not be necessary to use two different characters, and others saying that using two characters would cut down on human error.
There was also discussion of how diacritics are marked. USMARC and CANMARC do not handle them the same way, and actual practice, including LC's practice through the years, is mixed. There was also a discussion of the need for sorting and non-sorting guidelines. Such guidelines might need to vary by country and language used.
Jacquie Riley called for a straw vote. Most were in favor of implementing two new control characters and making the filing indicator obsolete for those fields. Most also favored using the control characters in all fields except the 0XX fields.
More discussion ensued about whether to try to amend the proposal to include diacritics or whether to deal with the immediate problem of filing indicators and deal with diacritics separately. Diane Hillmann moved to approve Proposal no. 98-16 in principle, but to defer a final vote to the 1999 Midwinter conference. Paul Weiss seconded the motion. There was considerable discussion about whether this was a proper way to proceed. When the vote came, the motion to approve Proposal no. 98-16 in principle, but to defer a final vote to the 1999 Midwinter conference was defeated, with 4 voting in favor, and 5 against. The proposal was referred back to ND/MSO for additional work.
Discussion Paper 108: Recording Language of Heading in USMARC Authority Records
Sally McCallum introduced the discussion. These issues were originally raised by an IFLA Working Group on Authority Data Elements. The group had noted that, especially when working with subject authority records, there was often confusion between the language of the catalog and the language of the heading. There was also some interest in achieving cost savings through the use of Model A of the DP.
Concern was expressed about situations where there are multiple languages in headings, as was illustrated by several examples, and about marking language rather than cataloging/thesaurus rules. Because headings are established and live in an environment set by the cataloging or thesaurus rules used to formulate the heading, the linked record approach used by the Canadians and others is a good method (Model B) in multilingual situations. While marking can be interpreted as being needed to distinguish a language, it is probably more accurately described as being needed to distinguish a thesaurus, with the thesaurus having a base language. NLC authority file can be viewed as an English-based thesaurus and a French-based thesaurus, with many records appropriate for either and so indicated at the record level.
Discussion turned to the best way to choose a default language. Would this be the language of the catalog or the language of the cataloging? There could also be the problem of conflicts with cross-references in other languages. There was some discussion of the experience of Swiss libraries in using catalogs in multiple languages.
Discussion of this paper will continue on the USMARC listserv.
Discussion Paper 109: Identifying Transliteration Schemes in USMARC formats
Sally McCallum introduced the discussion paper. This issue arose from CPSO's need to identify Pinyin and Wade-Giles transliterations in MARC records There was however some feeling that the need to identify transliteration schemes might exist more broadly. Because of lack of time, discussion moved quickly to the questions at the end of the paper. There were no comments on the first question. Opinion was divided on the second question: Is there a need to link corresponding transliteration-related fields, such as non-roman and transliterated roman data? There was no opinion on the third question. On the fourth question, the opinion was that the lowest level used for identification would be the field level. For the first part of the fifth question (What is the impact on cataloging efficiency?) the comment was made that there will be no impact unless systems can display non-Roman scripts. Discussion of this paper should be continued at the 1999 ALA Midwinter meeting in Philadelphia.
Discussion Paper 111: Alternate Graphics Without 880 in Bibliographic, Holdings, Authority, and Community Information Records
Sally McCallum introduced the paper, commenting that, although the discussion paper uses the term "non-roman," the appropriate name is "alternate graphics." Because time was getting short, discussion quickly turned to the questions at the end of the paper. On the first question, both Rich Greene and Joe Altimus commented that this change would have a very great impact on OCLC and RLG. Joan Aliprand elaborated that internally, RLG does not store information as an 880, but the 880 is important for validation when a record is being contributed to RLIN. In line with the second question, there was also discussion of how records with transliterations and alternate graphics would be handled. John Espley asked what will happen to systems that cannot handle alternate graphics if 880s are abolished and what was the impetus for considering doing so. Sally McCallum replied that using 880s is a Roman-centric solution, and may not be appropriate in the long term. Paul Weiss suggested that the problem could be solved by specifying the use of 880 more clearly. Joe Altimus commented that single fields that contain all scripts would be the most desirable model.
Due to time constraints, discussion will continue at the 1999 ALA Midwinter meeting in Philadelphia.
The next meeting will be Saturday January 30, Sunday January 31, and Monday February 1, 1999, at the usual times.
Time having expired, the Chair declared the meeting adjourned at 4:00 PM.Respectfully submitted,