ALA Midwinter - San Antonio, Tx. Jan. 14-18, 2000

8:30-12:30

Classification schedules were really developed a century ago based on even older literature. Until recently, when a printer broke, schedules were produced by Wordperfect 4.2. In 1995- began using classification schedule from automated authority records.

R,T, E-F will be available in a few months. LC hopes to publish schedules to meet demand.

LC Subject Cataloging Manual: Classification "F"

F50 - how to propose classification to LC

F195 - subarrangement within disciplines

Order of topics

F320 explains the different between time of publication vs. time of coverage

F475 - topic by subject that CPSO had to make a policy decision on, alphabetically

If proposing a class schedule that relates to any topic in the manual, make sure you read it

The beginning sections of the schedules usually have standard captions that have instructions in the SHM: Classification - read them; do NOT propose class numbers in this area

P.13 of handout - TH4808 - current practice would add "General works" before "History"

When making a proposal, we have to be aware of indentation

Cutters or cutter range A-Z are now in the column of class numbers when it is part of the class, but if a specific topic has a cutter with caption A-Z, it will be indented under the caption

Proposals for A-Z are the easiest to get approved, provided you are clear where on the page it belongs.

HM--- HM was rewritten; HN is large country-wide general topics with weird special topics; HQ gets more detailed : proceeds through life (youth, aging, death) and lots of stuff on women because they were the problem issue in the man-ruled world of the early 1900's in the US; HS is societies HT social problems, HX cures for the social problems in HV. If you wanted to add class for juvenile works about the aged, it might be 1061.3 (see 1073.3 on p. 15 of handout)

The numbers really don't mean anything - you have to keep the structure of the captions in order; the result may or may not result in whole numbers or decimals. You could even end up with a decimal number that looks like the tail end of a previous class number -- ignore that, look at the caption.

In the proposal, give the page number of the latest edition where the new class number will go - the one published by LC, not Gale.

[ ] means this is an anchor for the new one below it and whether it should be indented

(*) means change

If proposing a new heading that implies that the indention of another caption should change, that should also be anchored

If it says, eg. .N7-.N72 Norway, it is not supposed to be a complete list; you don't have to propose a new class number in order to use one on your own.

We don't propose index terms.

Filling in the proposal form:

Class letters = basic letters

Page = in LC's printed schedule

COOP = circle

Pattern = indicate if one is known

1. either a similar class structure in the schedule
2. SHM: Classification F xxx that is appropriate

"better" - only LC does this; it indicates a better way to indent and how many books in LC would need to be changed

Can pub class number, decimal, and cutter in Class number column because that is the new style

Captions go in captions

Do enter the work cataloged - author and title

COOP library code is the NUC codes, not OCLC

Web page instructions are a brief version of F50 in the manual. Please read the manual; if it's a really big proposal, talk to you r liaison. Note these changes to web instruction: 1) anchor is both class and caption, 2) use whole number or whole number and decimal that has text in the caption to position the new entry

P. 13 handout, ex: if proposing 4810 you would have to indicate if aligning with "By country" or "History and description" or "General works"

Align with History and desc

TH

4808 [History and description]

…

4810

OR

4810

4811 [General works]

----------

Align with By country

TH

[4809] By country

4810. By state

First think about your topic and where it fits, then worry about the number

P. 23 the pattern numbers and decimals are not the same as the proposal -- that doesn't matter; what matters is the phrase "study and teaching. Research" precedes "Collections and preservation" and those exact phrases were used

P. 24 leave space for actual number for cutter numeral to be filled in at LC

Do one proposal per page, unless what you're prosing affects a distinct class number elsewhere that also needs modifying

P. 25 (*) only give the new proposed caption

If you have more information, please include it - an abstract or explanatory note, or include information in an email to your SACO liaison

053 on SACO, use:

1. in case there is a class number that already exists that fits your subject being proposed
2. if you ALSO submit a classification proposal form -- if you don't need a class number, don't fill it in

Caption changes

• huge changes are discouraged
• if subject heading and caption don't match, LC encourages us to submit it because it will work better in Classification Plus -- although they are not eager to get author names in P's just because it is now initials and words spelled out in ( ) in authority records
• also don't bother with name of country proposals; they don't do it at LC
• encourage GOOD Cf. Notes to help find correct number

2:00-4:00

Primarily need to do appropriate and sufficient research but not exhaustive; make sure we follow Subject Heading Manual: Subject Headings

008/06 "i" or hyphen or "fill character" : the choices to fill in on the web - only fill in "Not" for things that might be through of as geographic subdivisions, but the decision is to NOT subdivide

053 - corresponding class schedule; this must be explicit. If it says … A-Z, don't try filling in anything, leave it empty

H187 - use first time you find a concept

H290-360 - principles like don't use articles, or use English form, or arrange 4xx alphabetically

4Xx needs 670 supporting documentation for variants or alternatives; "rearranged" only if appropriate because rules may prescribe it

If you have a subject and you need to qualify it, use as qualifier the city; if outside a city, use states/countries other than US …; county in state if more than one in a state

Note: Great Fire, London, England, 1666

5XX Fires #z England [do not add London]

If the subject could be a 1XX or 7XX, it is NOT a 6XX SACO proposal; i.e. can the "thing' in question write or create something? A dance troupe can create a dance; a building that you dance in, can't

If a corporate body inhabits a building, the same heading is used for both; if a corporate boy doesn't inhabit a building, it is a 6XX

5XX - just about always give a BT heading; if it doesn't exist, it must also be created

5XX - related - very uncommon for buildings or events; same is true for narrower

670 - work cataloged -- if the title or subtitle, has the name you're trying to put in the 1XX, you don't need to put something in ( )

If the only source your have for establishing a heading is the work you have (670), it is helpful to include additional information from the source

You do research to indicate a consensus of usage in relevant sources

If citing a dictionary or work logically arranged so it is easy for someone to find the word(s) in their copy of the resources, you don't need to give a page number; it if isn't obvious, give a page number

If you need to consult, Yahoo, etc.: AltaVista search, Jan. 15, 2000 #b (800 sources found ____; 5 sources found ___); or #b (usage in ___ documents retrieved: xxxxxxx; xxxxxx). (see also p. 19 of handout) In the former example, the 800 would be the 1xx and the form found in 5 would be the 4xx

952 - list things you want the liaison to check

LC pattern or SCM memo - only use if it's unclear otherwise or it's the only way you can show why your picked the heading and references you did

Events have to be noteworthy, in a specific place, and interval of time; floods, storms, etc. are just topic #z __

Whatever events aren't planned, end up as general subjects; although some may eventually become known as something and can be changed

LC doesn't go back and fix bibliographic records; they only fix authority records unless they have a work that causes them to verify it and find that it needs changing; if we find a situation, report it to LC

Why doesn't LC put in 4XX as earlier headings, especially if it is the same as the old heading but is changing between 15x to 11x -- they just don't!

Dates at the end of an event, do single or span; if several occur in one year, then add month spelled out, then day

P.11 BT Hurricanes - do NOT add geographic subdivision in BT

For aircraft accidents, LC does not do #z (some would be hard or vague to figure out)

P. 11 events that have names, use capital letters

P. 13 subdivision with date, use #y

Events tend to be phrases, but historic ones, like wars tend to be subdivisions unless … see p. 14 .. so how do we decide? Look for similar patterns

P. 16 #1 use current place name as qualifier

P. 17 to establish battle of a place, the place must also be established; if the battle is known with an old town name, use it in the battle name, but qualify with the current place name

P. 20 Buildings -- * MARC tags section

P. 21 Structures outside a city -- if more than one building with the same name in same state, use (county, state); also include the qualifier even if the name is in part of the first part of the name

If a building changes its name, use the current name and add 4XX (not like the corporate name in NACO where each one gets a new record)

Function may be brought out about buildings, but not construction materials

1:30-5:30

(came in very late due to conflict) (picked up handouts for earlier presentations)

Report of experience with CORC in Mexico, Jesus Lau: goals were to satisfy user demand, integrate web resources in the OPAC, strengthen the OPAC as the main library retrieval tool, fill in gaps in the national catalog, develop a metadata experience.

Comments/questions at end:

• Dublin Core isn't really so simple, and it can be made very complex; you do have a choice what you want to do.

• Do SOMETHING; plan to revise it once you figure it out

• CORC is an effort to find a way to do something quickly, if that is your choice

• How do we achieve interoperability if you've done a simple Dublin Core and want to crosswalk to MARC21?

• Sites you have to register or subscribe to, do you catalog? Karen Calhoun - originally they decided not, but sometimes they just can't find any other way to access something they want to catalog

• XML/RDF to authorize names? - any cooperation with other professions? There is a working Group (Dublin Core) that may start work on names; music publishers are becoming interested in name authorities because of royalties.

2 meetings, 1 business Friday 6:00-8:00 PM and 1 forum Sat. 2:00-4:00 PM

Business meeting.

IFLA - Lynn Howarth has indicated their metadata program in 2000 will include our report

Our report when approved by SAC can be on our web as "draft"; CCS Exec at Midwinter still has to meet

OCLC report, Eric Childress

• July 2000 is planned for bringing CORC into production and OCLC will charge similar to other services

• It will be integrated with WorldCat and synchronized with MARC21

• Records will be transformed into bib level 3 if just Dublin Core data available

• They would charge for the auto-Dewey unless it isn't working well enough; then they may remove it completely

• OCLC will introduce member-input authority records with mechanisms to convert to NACO; the record could be updated to NACO standards and keep the same ARN#

Program planning for annual 2001 must be done by April 1. Shelby Harken chair, also: Shannon Hoffman, Rebecca Culbertson, Sandy Roe.

Program thoughts:

• FAST

• CORC implementation

• Our report (Lois Mai Chan)

• Other metadata types - are they using it?

• Classification: make DD work better or use LCC as subject

• Foreign languages

• Visual resources - form vs. real subject

• XML/RDF implementation including subject

Overview of various metadata schemes (Robert Pillow, Becky Culbertson, Shannon Hoffman, Greg Wool)

• Book of interest? Getting mileage out of metadata

• Many schemes are designed for very specific communities

• TEI (20 elements) - quite old and overlaps (sort of) with Dublin Core but it doesn't have subject in its header because it is like an electronic title page

• DESIRE project - (how to compare metadata schemes; program thoughts)

• Environment of use
• Documentation
• Constituency of use
• Ease of creation
• Progress towards international standards, etc
• Format issues
• Designation
• Content
• Basic descriptive elements
• Subject description
• URI
• Resource format and technical characteristics
• Host
• Provenance, source
• Terms of availability

• Rules for construction of these elements

• Encoding

• Multi-lingual issues

• Ability to represent relations between objects

• EDNA - a slight expansion of Dublin Core with 9 more elements for them only, including their classification scheme, reviewer, date to reassess

Program thoughts: might be a place in the program to do a summary of various metadata schemes and their subject analysis

Program thoughts:

• in business marketing, you’re trying to hit as many people as you can;

• subject analysis is trying to focus use as being relevant to what you are looking for (it seems it would be even better marketing to target appropriate possible customers)

• various metadata have developed to meet specific needs, EAD, etc. whereas Dublin Core is very general supposedly applicable to all

• where does our report or subject analysis or subject assignment fit in specialized needs?

• is it useful or does it clutter the issue if many subject thesauri are being used?

• Where does interoperability of thesauri fit into the picture? Does it help? Can it be done?

• Is it fair to view Dublin Core as a "switching yard" to other metadata types or mapping through extensions

• Would it be possible to generate a report: on as many schemas as we can find, see how they handle subjects, what are the commonalities and what are the differences in subject construction or assignment and therefore is mapping or crosswalking possible?

• Could we have the above report and our completed report at the program? A committee to review varying metadata -- could be a handout

• What about geospatial, visual (art), etc.? -- we could get way out of our scope of work or so deep we'd lose people

• FGDC, GILS, DC, EAD, MINI, CIMI, etc. -- some don't do much with subjects - let's exclude them - share list by Mar. 1, draft due June 15

• IFLA site lists a bunch of metadata schemes and projects using them (also Lynn Howarth has a page)

Forum on classification - mid May draft for Annual forum 2000

SAC Subcommittee on Metadata and Subject Analysis Discussion Forum (SATURDAY 2-4)

Diane Dates Casey announced that our report, Subject Data in the Metadata Record has been approved by the Subject Analysis Committee

Lois Mai Chan (University of Kentucky; main writer of report) - presentation: rationale behind our recommendations

The subcommittee's task has been to identify and study major issues surrounding the use of metadata in the subject analysis and classification of digital resources; and provide discussion forums and programs.

• keyword (free-text) only

• sounds easy but! - lack of synonym control

• lack of homograph control

• lack of indication of term relationships

• schemes

• existing schemes - are they suitable for the web?
• Adapt or modify existing schemes?
• Develop new scheme? Who would do that?
• Controlled

• is one suited to everyone's needs? Which one?

• Generalized vs. specialized?

• Multiple vocabularies - how do we harmonize?

• controlled vocabularies - structure issues

• semantics

• term specificity - what level of specificity

• synonyms and homographs

• term relationships

• syntax

• full-string

• single concepts

• what's best in web

• controlled vocabularies - application issues

• consistency
• summarization vs. exhaustive indexing
• placement of non-topic: time, space, form
• pre-coordination vs. post-coordination - application syntax; putting terms together

• possible strategies for implementing controlled vocabularies

• define functional requirements of subject data in the metadata record

• new vocabulary

• examine, determine suitability of exiting vocabularies

• strategies

• simple to apply and comprehend
• intuitive so sophisticated training is not necessary
• capable of simple to complex implementation
• ?
• be appropriate to specific domain or subject discipline
• interoperable across disciplines

• role of LCSH in OPAC

• LCSH is a rich vocabulary covering all subject areas
• Synonym and homograph control is included
• Pre-coordinated precision in retrieval
• Facilitates browsing
• Defacto international standard

• advantages

• Because of rich vocabulary, synonym and homograph and term relationships, LCSH could continue to be useful and contribute to subject access on Web
• Would integrate well in OPAC where MARC records reside

• limitations

• because its complex, it is hard to train
• costly to maintain
• not compatible with other subject thesauri
• not in Web search engines

• other issues

• we do have trained catalogers
• we have systems with index searching and browsing
• we have sophisticated online display
• we have to acknowledge semantics and syntax
• without sacrificing rich vocabulary, can we separate form, time, place?
• We need to design a flexible syntax from most complex to simplest allowing for varied implementation

• pre-coordination of ideas

• so if fully trained people are available, they can do it

• post-coordination

• simple to use
• index display simple authority control easy
• easy to use
• interoperable with other vocabularies

• given the lack of trained personnel, providing for post-coordination allows wider use
• issues: harmonization, feasibility of designing a broad metathesaurus or adding to the medical metathesaurus
• we need to provide and design displays of controlled vocabularies or investigation of such

• existing vs. new
• should class number or only caption be included?
• How do you handle cross classification -- when several class numbers would apply
• Need to study feasibility of classification display data through online display and interface
• Auto-mapping of text to class schemes

• considered desirability of controlled vocabulary and classification data
• considered semantics, syntax, relationships
• over-riding principles are the same as Dublin Core

• simplicity
• semantic interoperability
• flexibility

• distribution of effort for subject analysis: Information Professionals ---- Computer ---- metadata creators including authors

• information professionals

• vocabulary maintenance
• policies
• preparation of application guidelines
• software and system design for storing and displaying vocabularies

• computer - process
• metadata creators

• information professionals but not trained catalogers
• web masters
• authors of web resources

Diane Dates Casey's summary

• Combination of keyword and controlled vocabulary
• Use of multiple vocabularies can be accommodated, LCSH, Sears with or without modification
• LCSH can be complimented with other vocabularies
• Synonyms to be handled by vocabulary or with record
• Displays should be developed to show hierarchy and sideways relationships
• Should be able to be as simple or as complex to apply as desired
• Recommends simplified syntax
• Harmonization of multiple vocabularies or expansion of the medical Meta-Thesaurus
• Classification should be used by those trained
• Display classification for use by others
• Multiple class numbers should be allowed with designated of code
• Include in Subject Element both controlled and free-text to assist retrieval
• Criteria for selection of a vocabulary depends on subject content of database
• Use most specific subject of chose thesaurus
• Apply consistently
• Name of subject code and date assigned should be indicated
• Full LCSH encouraged by catalogers
• In Dublin Core, subject aspects can be included in type, coverage … topic, place, period, language
• Deconstructed subject should be so indicated
• Classification - base on subject domain and users being served
• Use classification at most specific level; include notation, but not cutters or captions
• Auto keyword indexing of text in document
• Auto classification generated based on resource
• Compare to similar items for subject generation
• Auto mapping from free text to classification of controlled vocabulary

Eric Childress - OCLC

Subject metadata equation; Or why SAC + LCSH + DC + CORC = FAST

Today's discussion: (see handout)

Subject access in OCLC

OCLC CORC

Capture subject metadata

Synchronization of Dublin Core and MARC

Overview of FAST (Faceted Application of Subject Terminology)

• Dublin Core: 15 elements, designed to support interoperability, and translated into 26 languages

• SAC related elements: Title, Subject*, Coverage*, Type*, Format Source, Description, Language
• Dublin Core is intended to be simple to apply but can be implemented in very complex applications
• Identifier = token in Dublin Core
• DC.Coverage definition assumes use of words, but could use encoded data
• DC.Type - there is new proposed list of format types on the DC page
• Namespace - can be a pointer in a bib records to a controlled vocabulary, e.g. ARN# and display the subject to the user
• Scorpion - Dewey call number proposal
• Wordsmith - text analysis to provide keyword
• If you use Scorpion, it will suggest subjects that have been identified as being linked to that Dewey class
• OCLC looking at adding LCSH Children's
• Looking at some basic guidelines for use within CORC

FAST (Faceted Application of Subject Terminology)

• Research project - not ready for prime time yet
• Requirements based on SAC report
• Also look at Arlie House recommendations
• FAST is derived from LCSH - it facets what's already there; it is:

• Post-coordinated
• Designed for web environment
• Enables broad range of users
• Can co-exist with each other

• If you enter LCSH - you can redo into FAST, but not the other way
• MARC tagging slide - draft (see handout)
• OCLC is looking at running a program looking at form words in a table of form headings and converting to #v; form subdivisions will be removed in put in form headings

• Going to map back to highest Geographic Area Code (GAC)
• Turned 20^th century into 1900-1999
• #2 fast is a valid subject source
• Implementation issues

• Forms (type) - hard to decide which
• Geographic - hard to do things like "sun", "moon", "submarine trench"
• Topicals - hard to do fictitious characters
• Uniform titles - it has been decided this is not a DC qualifier; may need to define DC library elements
• Need help on DC.Type
• What about metropolitan areas and cities that cover 2 GAC's, repeat?
• Need to keep FAST in synch with LCSH
• River will be at highest level GAC - (Red River of the North would be North America because it crosses countries)
• Coverage - qualifier for jurisdiction?
• 650-655 or do all as 654?
• XML - are we ready - not yet
• We are moving closer to UNIMARC/XML/RDF model
• OCLC will communicate with all system vendors with they start using it
• How to deal with subject headings that are really form

(I missed some of the meeting due to conflict with SAC Committee)

Next meeting go through ALA to schedule room. Noon-5 July 7

9:30-12:30

Jean Hirons - report on seriality

• Bib lvl, 260, latest - goal is Sept. 2000
• Expanding chapter 12 for continuing serial and non-serial, websites, successively & integrated
• Plan discussion paper or proposal by annual
• ISBD (CR) - new name
• Define "I" or redefine "s" - need to revisit
• JSC approved changing AACR2 to be reorganized by areas of description; until then, changes will be changed in chpt. 12
• Whether description is based on earliest or latest, 260 needs to be repeatable
• "integrating entry" is a new name
• 500 note for latest issue consulted will be used

2000-1 proposal: Definition of subfield #z (numbering scheme) in 853-55 holdings (captions and patterns)

• numerals - Z39.71 for 863 says convert all numerals to Arabic; the 853 is used in predicting what you expect to receive or collapse functions
• Diane H. felt #2 option might allow display of correct format, while functionality could be following Z39.71 and use Arabic numerals for prediction and checkin
• Issues were brought up about a number of foreign languages; our Arabic numbers are not what Arabic people use
• Unicode has separate numbers for Hebrew; Unicode should know which direction the number reads because it knows which language goes left-right or right-left - that should not be an issue here
• Alpha-numeric, ordinal, symbolic, foreign numerals need to be considered; option 2 looks better, but too many questions; could it be fixed-length to the codes?? - maybe not necessary
• Bring back at Annual

2000-3 proposal: Definition of subfield #2 (Source of term) in field 583 of MARC 21 bib and holdgs

• Possible use: archival (preservation, retention), rarebook, collection development, [sounds like MULS];
• NLM is looking at it for recording retention for electronic stuff in the 853
• Assumption is #2 applies to all subfields so the field would be repeatable (because #2 would not be) -- not sure we can say right now because dates might be ISO but the terms in #a, l etc. might be rarebook thesauri
• Approved; 583 repeatable

Discussion Paper 120: Community information format integration

• Pro: might make it easier to integrate in OPAC; save documentation
• Con: processing would require major changes to the leader, change requirement to have 245 in bib format, authority work is implied
• Display - we have experience integrating authority records into OPAC so it could or could not integrate if we leave them as separate formats
• No matter what, you'd still have to make a decision on CI vs bib or particular leader
• Decided to discontinue discussion

2000-4 proposal: Anonymous attribution information in personal name headings

• #g or #j would be part of the heading, could be indexed and searched - much email discussion - see folder

• Option 2 passed with #j; need to clarify #g is very miscellaneous; #j is information where some information is known about anonymous people

Z39.50 attribute sets (Lennie Stovel)

• Attribute set. Attributes define which field, word or phrase, truncated or not, type things can be defined

• Now there are separate attribute sets -- e.g. GILS has its own attribute set because they felt 1^st version was not good enough
• New attribute architecture -- includes 14 attribute sets; now certain across-the-board attributes have been defined
• Utility set: another set very similar to Dublin Core
• With version 3 of Z39.50, you can use as many attribute sets you want to use in your query
• It is now expected that each set should be developed by each special group -- e.g. bibs by librarians; Liv Holm is the person on the ISO international working group
• Intended to be used with "good" cataloging of any format so display on the other end can be displayed in a general MARC or MARC-like standard

Library of Congress update - Sally

• Publication of MARC21 Authority soon; followed by others
• Specification will be on web in July
• Language codes - working on new edition after Feb. - within 6 mos.; LC will be able to implement new codes including the ISO 3-character code
• Experimenting with making classification available on the web

(missed beginning because at OCLC breakfast) 8:30-11:00AM

Presentations of multi-database search systems (see handouts)

• Getty has in OPAC both MARC records and things like 100,000 records for auction catalogs; the authority system for the auction catalogs is different than for books so the entries HAD to be different than entries or cross-references on "real" authority records because you don't want to auto-flip records wrongly
• OCLC FirstSearch - allows users to pick multiple databases before you start your search; you can also search the web -- get to NetFirst which gets your CORC records
• Pharos (Calif.) - 23 campus OPAC's and e-journals they subscribe to, Infotrack, and other indexing databases and ILL requests; includes both free and subscription databases in the list you can pick from; can only do a keyword search; in future hope to provide for limiting; right now can only use a thesauri for searching once you go into a particular database; all must be Z39.50 databases
• Searchlight -- has broad categories - OPAC and free database to outside users; keyword search; once you search, shows choices at top for reference sources, or web resources or books, or at bottom it shows number of hits in each specific database; it still needs to be able to limit - future would hope to limit by keywords - title, abstracts, and subjects if available; if you pick PubMed, Psychinfo, ERIC wizard, once you do keyword search, also show the terms in their thesauri and allow user to click on their thesauri to initiate search; http://www.csun.edu/mwoodley/Protypes/index.htm
• We still are not much better in the databases we've developed than general AltaVista etc because users are just putting in keywords and not getting much precision which cataloged records provide; no one seems to be trying to use a metathesaurus or map 2 or more thesauri (controlled vocabulary)
• Some databases, when you pick them, may begin to lead you to redo searches; for others, there is nothing; there's nothing at the beginning of any of the big pages to explain how to search effectively
• It also depends on how the vendor at the other end defined the fields in their records, so you could select a single database that you get through different vendors and get different results

Dublin Core - Stu Weibel and Rebecca Gunther

• 4^th Dublin Core meeting started the idea of qualifiers to explain each element; qualifiers can have encoding schemes that define how that element qualifier should be read, particularly by machine for display or invoking additional processing some way - i.e. go out and do authority control on a keyword in LCSH
• after ballot, Feb. 11, the set of qualifiers will be approved; expect future growth in qualifiers
• requires new software for registry data that will be check or work with data in Dublin Core records
• Dublin Core -libraries is a new working group you can join
• There will be a registry working group. Check OCLC's DC pages for mailbase groups
• Some qualifiers in CORC now may end up changing after balloting; newt, will need to decide if some things don't map, do we shoe horn them in or add a special set (domain specific); also do we decide in our community we may never use a Dublin Core qualifier
• Uniform title will not be part of general Dublin Core because it's so rule-controlled; may need to be added as domain-specific
• We need to make sure that registries exist for what we need to use; e.g. LCSH; that also infers that we need access to the actual records in the "file" defined by the registry - e.g. a URL to LCSH; another ex.: a need exists for a list of types - similar to 06 types but probably need to be more specific -- is there one out there we can use or do we need to develop one? E.g. a=text -- maybe "text, journal", "text, thesis"
• DC elements: source vs. relation. Don't use "source" unless you/re only using bare DC without qualifier; "relation" will become better with qualifiers
• Changes to AACR and MARC21 - ISBD(ER) - is now (CR)
• JSC has made comments on proposals; Brian Schottlaender is leading an ALA group discussing how to change AACR2 for metadata
• There are many metadata schemes, but Dublin Core is one we are beginning to know something about; 2 elements do go from Dublin Core to MARC well: date (date last modified means for the metadata, not the bib record); and relation; they will be doing a paper on these issues once DC qualifier balloting is done
• Crosswalk will be updated after Dublin Core balloting for MARC/GILS/DC

Final report of this task force will be on the metamarda list.

9:30-11:00 AM

(heard only last presenter because I was at Metadata meeting)

Silverplatter discussion [email for a copy of presentation cmeyer@silverplatter.com]

• Annual fee for unlimited user site license
• Discounts available for consortia
• More publishers are beginning to disengage the print/online combination for subscriptions
• Archiving
• Perpetual access - you've bought it once, ought to have it forever; for a business (SilverPlatter) this is a liability
• Digital preservation - this is more refreshing the digital to current technology
• Migrating data files to new formats
• Depositories
• SilverPlatter recognizes the problems and sees itself as a link to retrospective archives when possible; cumulative bibliographic databases are more valuable with links to historical content - i.e. archives at publishers pre-electronic version

Proquest does not notify libraries of dropped titles; you can look it up; they try to replace the lost title with access to a similar title. They focus more on keep a target number of journal titles in a discipline

Project Muse keeps usage statistics and is starting to archive

Proquest keeps statistics

Silverlink does not keep statistics

12:00-1:00

EBSCO-host

• 3000 journals indexed and abstracted
• 1240 full-text with 300 full text back to 1990; 1000 journals with images
• they determine titles to include in "academic" based on what academic libraries subscribe to
• Academic elite includes a company listing; it has 1700 peer-reviewed journals that are indexed
• In spring, they will allow soft linking from EBSCO-Host to other databases you also subscribe to - links to EBSCO online so if you subscribe to print and get online free, you will get through EBSCO-host to full-text of those journals
• Has HealthSource [not sure how this compares to CINHAL)
• Remote access is possible via homepage
• Updated daily
• Can set up to link to local holdings
• Can search multiple databases at one time - does de-dup if you want; precision of results is not good
• Keeps a search history
• Coming soon - persistent linking - faculty can use to save a link for students to use; page composer - can create a web page with links to EBSCO, image library, map library (mapquest)
• EBSCO Admin customize for us - local holdings and notes; authentication - IP, barcode, whatever; collection development statistics - can run reports;
• In next few months you will be able to email PDF's from EBSCO-host
• You can get to EBSO online full-text from initial search list
• Company directory - can search free and it shows SIC code, but it costs to get Dunn market report
• MARC files will be available to select from to obtain bib records (only for EBSCO-host academic elite the last I heard)

(missed the beginning) 1:00-3:00 PM

CORC

• Explore automated tools, cooperation among libraries and librarians, use a model we know works
• Tool for creating bibliographic records for electronic documents; can be done because machine can scan digitized document for text and suggest possible keywords and Dewey; it assigns MARC and/or Dublin Core elements
• Bibliographic records are linked to authority records so if an authority record changes, the bib will be updated

Vision for the future

• Library can be at the center - make libraries' web sites the portal of choice
• Can include your collection plus international and local resources selected for your patrons
• Can provide remote access or self-service mode
• Put in your OPAC because patron has no clue what journals you have, which databases they're in, whether they're free or licensed
• WorldCat is going to have to include more than MARC - Dublin Core, EAD, etc.; also other kinds of data like book cover images; need links to subject and name headings; links to other agencies that libraries work with - vendors, ill, TOC, biography of author, book review, etc.
• How do we fit in to the "earth's biggest library" idea? -- we should view it as building the best library with local knowledge, flavor, expertise

Future functionality

• Validation on the MARC form of the Dublin Core would be at MARC level 3 but could be upgraded

• Pricing - libraries see it similar to bib records; transaction-based for first year; may try new models later - maybe flat fee; if you are a fixed fee library, it won't go up the first year; same pricing includes credits for original cataloging

• If you're already an OCLC member with an authorization, you won't need a new authorization

• Once you set holdings for a CORC record, you could get bib notification service for alerts to changes - URL's and authority records

• Expect CORC records available for searching in First Search at no extra charge
• NetFirst will also contain these records - in all, it is probably going to become one database; but NetFirst is a separate database some libraries buy alone
• Hope to continue including old-fashioned-ly cataloged items to be brought over to new database; future is hop to provide web-based cataloging rather than the old-fashioned kind - in the meantime they'd like us to keep experimenting
• They are still working on URL checkers - they also haven't decided how to designate on a bib that a URL no longer works
• Audience - person asked for linking with LCC and MeSH and other thesauri or class systems
• CORC & CONSER - working with those groups - right now the CONSER library MUST still use old-fashioned OCLC; hope by summer to be able to do CORC

Exhibits - stopped at Ex Libris both - got demo for 1 1/2 hours

6:00-8:00 PM

7:00-9:00 PM

PCC - 180,000 new and changed NAR's and SAR's; 1,413,240 total authority records

SACO - 2000 SACO records contributed; SCCTP training underway; SACO web proposal form is for BIBCO libraries only for a test, but should be available for the rest of us late

Task Group for Aggregator Databases will continue for 2 years; new info will soon be on web

Automated classification - Gary Strawn has been been working on it

38 BIBCO libraries entered a total of 297,795 bibs, 19,638 of which were Core records; GPO is becoming a BIBCO library; if default on series authority records says to catalog as set or says to catalog as series and you are a BIBCO library, you can still locally choose whether to cat sep or as a series

CATME - some significant interest has been expressed in creation and real-time update of authority records

They are exploring web access to the NACO manual

CONSER update - Jean Hirons

• Several new members are joining
• SCCTP training is working great
• Going to develop a holdings course
• Publication patterns - will conduct a 2-year experiment using the 891 in OCLC to hold patterns and examples; will see the database with Harvard's database; documentation will be on web; non-CONSER libraries may use but probably won't see it until April; trying to create standardized checkin data and share patterns
• 510 data in CONSER records - JACC med library at Yale - JAKE.med.yale.edu - storing 510 data - not sure where this will go - could remove 510 in bibs and replace with a URL to Yale
• AACR2 - preparing revised chapter 12 to include continuing entities, etc.; ISSN & ISBD(S) need to be harmonized

Joan Scuitima talked on need for fluidity in standards yet need to maintain and follow standards

• Cross-refs serve different needs and have different uses
• Not intended to help those who do typos, pick wrong database, etc.
• Local systems often have problems references can't solve
• There are 17 recommendations: 5 AACR2 and 12 LCRI's, one being varying spelling and form of name that should have a cross-ref
• References to help those who don't know "our" rules - eliminate references concerning exclusions in certain languages
• Still need to address issues of non-Roman script
• Felt we would benefit from cross-references for former headings as we often see on legacy headings
• Make head of state-personal name be bi-directional
• References are often specific to a particular library, i.e. LC which makes no sense in libraries that don't have that title; need more latitude
• Drop #f in 400 for Bible
• Should be able to construct Dept. of …/…Dept. without proof
• Decided against tiny variations as cross-references - the big shared record needs to be usable by all

(missed due to conflict)

7:00-8:00 AM

• This is a distributed digitization project
• Sustainability and migration of digital objects - i.e. creating a digital archive
• Setting standards and doing training
• Started work on an NSF grant
• 2-3 museums, 2-3 libraries, and state libraries in each state received letters
• looking at libraries, museums, etc. for statewide projects already underway
• looking at an NSF grant for the region to store data for all at University of Utah unless that doesn't seem to be the right way to do it
• Feb. 11 - what do we need to be doing? Can a collaborative effort actually work and what should we embark on?
• The project is one of 6 ideas that came out of an earlier meeting
• Only 3 statewide initiatives were identified; 2 city public libraries had a project
• In Colorado, making sure access is available via Z39.50
• Asked historical societies and museums to come to Feb. 11 meeting
• Sara received an RFP from NSF -- looking at how high speed computing would move around the consortia, archive data, and make information available in rural areas
• Center for high-resolution to manage research material
• Also looking for faculty interested in western history who want to get involved
• Missed the last grant but want to be ready for the next grant - Nov.
• Could do projects on themes by pulling together archival stuff along the way in libraries: trails (e.g. Oregon Trail), Native Americans, etc.
• Amigos has a NEH imaging grant and staff with knowledge
• NEH is trying to set up digitization centers
• There are a number of standards or best-practice guidelines - we need to review where we are at
• What do we collaborate on? We need a theme, e.g. what built the West?
• Feb. 11 meeting

• What collections do we have in common - maybe a more detailed survey and what are the top 3 priorities
• Funding options - can range from $500,000 to $3,000,000
• Decide for next step, pooling various types of expertise: subject expertise, digitizing-scanning, etc.
• How to bring in more museums and historical societies and expand participation
• State information literacy codes - can we support this in some way for K-12? And life-long learning
• Create public awareness of the project - market - the idea is this is a project that users could be pointed to instead of people wondering if info is in OPAC's or on web and which library might be doing anything on Western history
• Looking at OCLC Site Search to be the searching interface
• Maybe we need a web page

9:30-12:00

They asked where ODIN was, what I was looking for. They showed me their system.

2:00-4:00

• Aggregator survey results

• 71% wanted bibliographic records in OPACs
• vendors need to be the source
• aggregating is never done - always updating
• prefer one online record in OPAC for all online titles

• Cal State Northridge loaded records available from EBSCO and are happy (Academic Elite, EBSCO-host)

• Ways to do analytics for databases

1. human - best version but takes too long
2. machine derived - use existing bibs and certain added fields
3. machine generated - no authority control, subjects, or call numbers
4. scripted creation of minimal records - with NO access points
5. index at aggregator's database - it is not in your OPAC even by title, much less
6. typically, 200-300 titles are logical to do human; more, prefer aggregated

• 773 is repeatable so one online record can be used; decided to use "s" instead of "b" bib level (except OCLC and OPAC's don’t understand it)

• distribution - 1^st you get one big set, then only records need change or deletion; OR delete everything, then load a whole new updated file

• How to handle "dropped" titles: there are 2 kinds, one is just closed; the other is, you no longer have any access

• Coverage must be available to users in bib for each 773, or summary holdings

• Title changes - we need notification from vendors

• Aggregator report has been accepted; now pursing more vendor cooperation; test loading and maintenance (e.g. close vs. drop needs to have a code that would cause the receiving system to handle properly); how to handle electronic books, mainstream bib standards; final report will be on the PCC page

--------------

Regina Reynolds report on ISSN rules

• Revision goals: harmonization with ISBD(S) and AACR2, have rules for new media, fewer major title changes, update ISSN manual with examples
• Need ISSN for: electronic journals, databases that are continuing -- just don't know about web sites
• What version of electronic journal should have new ISSN? PDF? Integrating?
• Should URL for ceased serial stay in ISSN record?
• Proposing title proper plus qualifier (IF necessary); qualifiers would be to add qualifier in ( ) dot; when initials start journal titles and full form is present, use full form
• There is significant disagreement on the qualifier; only US and Canada use authorized form in ( ) - the rest of the world uses whatever it says on the journal
• Only 30% of records have ( ), and of those only 10% have a body name
• ISSN manual is still using successive entry for the files
• First how-many words constitute title change? 3 vs. 5? First word change is nearly always major; disregard rearrangement of corporate body in title; if in doubt consider the change to be minor

-----------------

• JSC will be making final decisions in Sept. 2000
• For now, expanding chpt. 12 and 21 but will eventually re-write by areas; appendix will contain rules for major and minor changes
• Finite and continuing resources, also integrating resources, will be in chpt. 12
• Latest approach for integrating resources and call an integrating entry with new code - MARBI
• Chpt. 12 revisions

• Wording of serial to continuing resource
• Rules for integrating resources to be integrated with rules for successive by area
• New rules added where needed regarding changes; e.g. what to do if you decide you have a title change (since they haven't decided on 3 or 5 words, so for now it is still 5)
• New examples
• Sources of information

1. Successive: earliest issue
2. Integrating: latest iteration
3. E-journals-successive (at least for now)

• For e-serials use the most formal presentation of the title
• If you find typo - just fix it, don't use (sic)
• Don't use "welcome to"
• 246

• full form and initials
• 1.1E6
• statement of responsibility embedded

• use judgement (i.e. - don't need a 246 for each minute variation)
• designation

• if no designators on 1^st, use date
• if numbering goes back to the beginning, we can supply "new series"
• publisher - still need to use first one; asking MARBI to allow 2 260's

• 362 plus 500 Latest issue consulted: v. 30 (1999)

• ISBD(S) - will cover continuing resources -- now it is called (SBC (CR) serials and other continuing resources
• Serial is a condition of publication - not a format
• Serials continuing resources in any medium issued in succession of discrete parts, usually bearing numeric or chronological designation and usually having no predetermined conclusion
• IST = International Standard Title (proposed)

• IST for determination of major and minor changes
• IST would replace key title and uniform title in part (ISSN's)
• IST in 222 - most other countries use this field for linking

• Description

• Earliest vs. latest? Legitimate reasons for each
• Title proper and publisher are key issues particularly if bib is being used for claims

• Title changes

• First 5 words as far as ISBD is concerned
• Retain all current "minor" changes lists but it (the list) will be reviewed (singulars, plurals, etc.)
• Approved issuing body; words denoting type
• Slight variances in lists of things in a title, e.g. directory of Canada, US, & Mexico, then Directory of US, Canada, & Mexico

• Remember - this is still discussion!!!
• What happens to an OCLC record when ISSN folks change things
• What will uniform title end up looking like? Abbreviated key title will still exist
• These changes would be from this day forward and no going back
• IST - to distinguish: a) major vs. minor, from b) similar
• Our current uniform title 130 would be a 730 and the IST would be the 130 therefore there would be no 130 like "journal title (Online : JSTOR (Organization)
• IST is trying to be an international name that would be stable
• 130 would be for legal documents, etc. but not like what we are doing now
• Two groups are still disagreeing about the presence/absence of words like magazine
• After May, there will be drafts from each of the 3 groups (AACR, ISSN, IFLA) involved in the discussions; IFLA's will go out for world wide review

CONSER report

• Integrated Library System (ILS) at Library of Congress - still learning about holdings patterns, etc.
• LSSI will convert manual serial checkin data - staff of 24 working 2 shifts
• There is a proposal out there to make ISSN a URN to be able to type in http:// followed by ISSN; can get to a list that is in 776 form with URL's for electronic

CC:DA update

• Trying to get AACR more principle-oriented, not format oriented
• 0.24 is a problem; description must address content, carrier, type, relation; if descriptive rules provide for different things that conflict, can use both; description rules should be organized with general first, followed by specific rules - prefer to use most specific that can be applied and if not, work backwards.
• What to do with multiple formats in 1 item?
• JSC agreed to the introductory chapter, esp. those suggested by 0.24 report

MARBI update

• 2000-1 numbering scheme needs more work
• 2000-6 773 not to contain URL; other fields OK'd in the proposal