ALA Annual Convention

June 25-29, 1999

New Orleans

World Wide PALS Users Group Meeting

Attendees: Paul Meyer, Jason Christians, Richard Oberlin, Heather Blenkinsopp, Suteera Apichatabutra, Stephanie R. Bernard, Lynn Karen, Ellen Kotrba, Dorothy Liegal, Michael LaCroix, Robert L. Fallon, David Barton, Becky Bell, Teresa Edwards, John Christenson, John Stromquist, Leah Monger, Melissa Beck, Phil Williams, Pat Nichols, Roger Presley, Donna Gilliland.

Minutes and Treasurer's Report.

Gary called the meeting to order. Teresa Edwards moved adoption of the agenda. Richard Oberlin seconded; passed. Shelby Harken had distributed the minutes of the previous meeting. She noted that CalWest is a community college district, not a university. Ellen moved the minutes be adopted with that correction. Teresa Edwards seconded; passed. Shelby gave the Treasurer's Report. There is $3863.33 in the account with a bill expected for about $800 from MnSCU/PALS for room charges for the last ALA, which leaves a balance of approximately $3000 in our account. Ellen moved adoption of the Treasurer's report. Richard Oberlin seconded; passed.

The PALS listserv is currently at a site that is migrating away from PALS. Gary offered to set one up in South Dakota and subscribe us all.

Status Reports.

Atlanta University [Robert Fallon] They have had problems with DepCon. They are on ver. 13. By Sept. next year the Unisys contract expires. They have had a consultant come in. They need to clearly understand how long PALS will last in order to make decisions. The medical school recently signed with Endeavor.

Creighton [LaCroix] They continue to look for a new system. Any new system must have UNIX and Oracle. SIRSI came in at a medium price and much lower than Endeavor. A decision will be made soon. A number of other consortia in Nebraska have picked SIRSI. ExLibris seemed to be a good system especially for international consortia, but reserves and media booking was weak. All three vendors recommended LTI for authority cleanup. They hope to be up by July 2000.

Ferris State [Leah Monger] They are part of a consortia, Ferrisnet, which has lost some members and a new one joined; one went to SIRSI and one the Dynix Scholar. They are on ver. 13. They will be building a new building in a year. They hope to put out an RFP later in the summer. They plan to do weeding and inventory before migrating. They are using the OCLC Micro Enhancer even for acquisitions and use an 049 that says ON ORDER for the bibs they bring in.

CalWest [Richard Oberlin] ZPALS is up. RFP bid went out in March. They had demos before the RFP and none after. They had replies from III, DRA, Endeavor, and SIRSI. The IT people determined that DRA is not functional; III didn't met hardware specs; SIRSI didn't want to contract with bonding; Endeavor's price was reasonable. All of the over 100 community colleges in California were given $200,000 for new systems. They plan to migrate and be implemented by Jan. 2000. Most libraries are not using Serials or Acquisitions. Bids are out for moving to TCP/IP, changing registration to Oracle, etc.

ODIN [Tony Stukel] ODIN has hired a new person, Ginny Millette. They have decided to do IAC databases via the web. Because the whole file wasn't loaded, one actually gets broader coverage by using the web. They are working with the ND State Library to contract with OCLC for SiteSearch with the box housed and maintained by ODIN. An RFI for a new system was issued and responses were received. Now they are waiting since it appears Minnesota will continue with PALS for a while. They didn't ask the legislature for money and since it meets every 2 years, it won't be until the next biennium that they could request funding. They are not likely to do an RFP until the fall before the next legislature. They are adding new libraries. They are on ver. 13 and anxiously awaiting ver. 14.

SDLN [Gary Johnson] SDLN is on ver. 13. They will be adding K-12 libraries by the end of summer. They have 14 databases loaded; many are local. They are looking at expanding services, e.g. FirstSearch. ND & SD are looking at SiteSearch, perhaps jointly. Their Unisys contract expires in 2 months so they are looking at a Unisys ClearPath as it would be the most economical if PALS is truly going to last 5 years.

TDS [John Christenson] "In 1997 there was the flood, in 1998 the tornado, in 1999 the library planning task force." Most PALS Users Groups were NOT in favor of continuing negotiations with DRA TAOS, while the University of Minnesota wanted to continue negotiations. Despite all testimony against DRA for lack of functionality, the vote was in favor of continuing. Several changes will need to be made in TDS because some libraries will be breaking up (politics) so bib and item records will have to be re-done. Funding has not been good for school and public libraries.

WALDO [John Stromquist] WALDO has a new member. Now there is a member from all 5 New York boroughs. The new member felt it would be 35% cheaper to use WALDO which is using MnSCU. They are also talking with SUNY-Fredonia for a temporary stop-gap until SUNY gets a new system up and running. They will probably create a position for executive director. Becky has done training there since some libraries want to add modules. They were able to contract for a single consortial link to OCLC for cataloging and ILL and save lots of money. Hoovers Online is available to K-12. An EBSCO package deal is being worked on. Looking at shared archival storage facilities. They are happy with PALS and happy to wait and see what the market does. SUNY is looking at Ex Libris, Endeavor, and SIRSI.

Most consortia seem to be expanding, even MnSCU, especially to K-12.

South Africa [Paul Meyer] They purchased the license to PALS. It is growing in South Africa and the rest of Africa. The Unisys platform presents a problem but it can be adjusted to their specific needs. They have 244 sites, 3.46 million bib records, and 11.6 million barcodes. They have lots of different kinds of libraries, but mostly public. They serve government supported libraries. Ex Libris is in Western Cape and III in the north for academic libraries. They have lots of libraries to expand to; within a year there will be 8 PALS consortia, in some cases several operating from one processor. They are looking to the Internet for future development. They developed their own Windows OPAC interface using Visual Basic. They are interested in what MnSCU is doing with Windows PALSTAC. Libraries have lots of choices in how they use PALS; local events can be created with a "txt" file. They can create files of local programs, tourist information, etc. using a word processor. They have also written programs to do things like open a cash drawer and record payment transactions after CIRC has done PPF.

Report from MnSCU/PALS.

Administrative update, MNLINK, and System X (DRA) [David Barton]

Becky Bell is now head of total user support for PALS and some of what she was doing administratively, Dave is now doing. When reporting to PALS-HELP, as long as you get an email back with a number, the report has been logged and will be addressed. Billing will be done the same as last year; bills should be out by mid-July. The MNLINK gateway is up and running using OCLC SitSearch and it's working fine. Getting all the libraries connected is taking a while - four departments in Minnesota have to review the documents. By year 2000 patrons will be able to do their own ILL requests directly online. Dave reports to three vice-chancellors at MnSCU. He said they will support PALS for 5 years and can provide a written statement if needed.

DRA: the contract committee has not yet met with DRA. The Technical Group reviewed TAOS and the RFP very thoroughly and came back with a negative report. User Group reports recommended discontinuing negotiations and the Price-Waterhouse report had some high risk assessments. However, politics, in the end, supported continuing negotiations with DRA. The contract committee will start meeting with DRA July 8. A contract will not be signed until July 2000. Training staff etc. will take to June 2001; then 3 years to move all libraries in Minnesota. If issues come up in the contract that cannot be resolved, there may be no contract with DRA. Despite it all, the MNLINK project has become a "positive" in Minnesota politics to get money for library development.

PALS Development Plan.

The PALS system is pretty complete. Windows PALSTAC is probably the biggest priority, then maybe the 3M selfcheck interface. It is more "enhancements" that they will continue, rather than "development". They haven't tested the Gateway enough to see if they can "kill" it.

Version 14.

User support. The WWPALS calls report represents only about 50% of user support. The blue brochure describes user support. Each system's support staff have passwords and should use capital letters in the subject line. If a system is down, call. The new software they are installing will provide the option to submit requests via web; it is supposed to have a method of tracking requests which we, as users, could see.

Development. There are 2 lists: 1 6-month, 1 2-year (blue).

Version 14 will have the newest ISO/ILL profile changes for both the mainframe and the client. The December software will have a new PALS icon. The PALS holdings format can already be output in MARC Holdings Format; new development will also provide for display. (This was developed for the OCLC union listing project which OCLC has placed a moratorium on). The 3M selfcheck has been added to the list. PALS can already do patron authentication because of SiteSearch software; 3M changed their software a lot - can do TCP/IP connections.

The scanners used for inventory are no longer available because they were not Y2K compatible, but that doesn't matter to PALS; so MnSCU/PALS is investigating how to get a new scanner that we can use with PALS.

PALS is watching standards and putting them on the development schedule.

COBOL-extended: COBOL has limited areas within which to work. After ver. 14 goes out, they will start using extended COBOL so we'll have to contact Unisys about it to get it on our mainframe (universal compiler)

Directable ILL - this works with branches - "I want to place it 'here' and pick it up 'there'."

Make reports available to use with other software - need comma-delimited - this will be done.

Local validation will be possible to set up with the MARC Editor.

The ability to initiate a Z39.50 search from within PALS session will be done; the UNIX side is done already.

As soon as OCLC announces a date for the 5-character OCLC code, PALS will set its date to accommodate it.

Email needs to work. It is not working in ver. 13.

Version 14. [see handouts]

Branches that have different 3- or 4-letter codes in the 049 can be configured separately.

DSD in WebPALS for titles with long holdings, eventually dies. It has been changed to skip 30 of the same level holding and will hopefully go far enough to continue the DSD. If not, contact MnSCU/PALS.

Ctrl F1-F5 will allow 90 characters for MARC Editor.

CIRC, added 4,5,6 to be used to lock the address for that one patron so it will not be changed if you change the whole patron class from one address type to another.

To discharge items at a wrong branch when that IS really what you want to do, you must be logged on and do DCG <item> INTRANSIT

Booking module is for rooms, equipment, AV, etc. You need to have circulation running in the library. You have to set up booking classes and other policies. Draft documentation is on the PALS web page under Publications.

ILL notices will be included when you manually send other notices to a patron.

Binding has a few new options.

ACQ 90 - can control how many years to store on tape.

Because of some new features in ver. 14, new areas will have to be initialized. One must run OC-60 immediately to keep statistics from being corrupted.

New PALSTAC will soon be available. Vermont Views is being used for PALSTAC Windows development. It is due by Dec. or so ..., maybe after Y2K.

Archives.

Shelby proposed that WW PALS User Group records be deposited in the E.B. Robinson Department of Special Collections, Orin G. Libby Manuscript Collection at the University of North Dakota. The current 3 years of materials would be held by the Secretary with the oldest year being deposited annually, usually after annual ALA. Teresa Edwards so moved; Michael LaCroix seconded; passed.

Future WWPUG meetings.

Teresa suggested that as long as development continues, the group should continue. Meeting annually at midwinter was suggested. Lynn Karen moved that the next meeting be Friday night at 6PM at midwinter ALA. Tony seconded; passed. The question of what to do with funds in the checking account: Gary moved we deplete the treasury and then fund any future meetings with registration or other individual fee. Ellen seconded; passed.

Gary nominated Jon Christenson for vice-president/president-elect. Ellen seconded; passed with a unanimous vote. Shelby is still Secretary/Treasurer.

PALS Reunion.

Madelyn Valunas from Shippensburg submitted a written report, read by Gary.

Report from former PALS user Shippensburg University of Pennsylvania

Overall, we are happy with the product and vendor support. Shippensburg migrated from PALS version 9. We accepted our data conversion about January 5, 1998, and are in production on the Endeavor Voyager system with cataloging, acquisitions/serials, circulation/reserves, and OPAC. We were in immediate production with cataloging, OPAC, and circulation. Reserves, serials check-in, and acquisition were phased in as we manually recreated data that could not be converted. We are closing out our first year of acquisitions and thus have no experience with Voyager fiscal year roll over features.

We are using the WebVoyage interface for OPAC but also have the Windows client available on the library OPAC stations. We expect to migrate to version 98.1 in August and to install the media scheduling module also at that time. Five of the 14 universities plan to implement media scheduling.

What we liked in PALS that we still don’t have:

Dynamic transaction log statistics, transactions by hour and day, acquisition backlog, inventory, more canned reports; automatic collapse of barcodes/holdings when journal volume is bound. Also, reserves in PALS was easier to search.

Relevancy ranking, 856 linking, hot links to browse list from subject, author, call number, etc

Workflow has changed most in acquisitions and cataloging as a result of migrating to Voyager. From my perspective (however, I believe many who have been affected will agree) this has been a positive improvement. We have done cross training and have a better sense of workflow and relationships to the larger operation; i.e., teamwork has developed. With PALS cataloging did all the bib record selection that, with a few exceptions, is now being done in acquisitions.

Serials presented the most work because we had to recreate check-in records and publishing patterns in the new system. I am not aware of anything that might have been done to avoid or alleviate this. Data also had to be recreated for reserves and vendor files.

I believe the vendor has met the requirements set forth in our Pennsylvania State System of Higher Education contract. There were some misunderstandings at some local sites but that was not the vendor’s fault. With the contract negotiated at the System level there was no collective knowledge about each of our individual systems.

Windows 95 training was required for all and was the responsibility of the local site. The vendor provided ‘train-the-trainer’ training, 4 days of functional training at regional sites, and 2 days at each campus. The latter was to be specified by the local library in terms of what and when.

Pat Nichols reported on DRA at the University of Manitoba. They are no longer a consortia; they all merged. They have 22,000 students, 2 million volumes, 9000 journals. They were on PALS 1989-1995. They brought their system up quickly, partly due to outside constraints: Mar-July, OPAC, CIRC, Cat; Feb - Serials, Apr. - Acq. Classic DRA is what they have, not TAOS. The interface is either menu-based or a web OPAC. They still don't have online statistics, particularly public access statistics. Other statistics are not a problem to get, but they are not online. They miss the ability to automatically create item records for serials at checkin; they have to set up the cataloging module for serial checkin to create item records or create on-the-fly records. There is no summary holdings display in the OPAC, although there is in Serials. The biggest improvement was OPAC interface (compared to old PALS). Web cat is not good for reserves or patron information. The next version is supposed to have that as well as hotlinks. Immediate indexing except for keyword - that is batch indexing. They didn't use PALS for Acquisitions because currency conversion wasn't available then. They load payment information into the university accounting system. Serials prediction is the most difficult - prediction, etc. is not compatible.

Inconsistencies in the call number caused problems in the item records; would have been able to avoid some problems if they had had more time than 138 days. They would have been able to get call numbers from bibs instead of item records, but DRA doesn't look at location. They feel DRA only made 2 major conversion errors. Training was train-the-trainer; in general, it worked well. Generally they are satisfied; they have had little down time. Workflow changes were most drastic for Acquisitions. All item record creation is done by Cataloging staff only. They ran the two systems side-by-side for two months.

TAOS? The story keeps dragging on; they keep getting told about 18 months. A new cataloging module is supposed to work in both classic DRA and TAOS but they haven't seen it. DRA support for Classic is very good. The main problem is the report writer which is not a DRA product. Delays in TAOS seem to result from DRA repeatedly changing the approach to writing the new TAOS software.

Phil Williams reported on Georgia. All state academic libraries will be on Endeavor. Phil gave the history of PALS across Georgia and Galileo and cooperative purchasing of serials. They issued and RFP and tried to keep if from being tainted by personalities. They had libraries using DRA, NOTIS, SIRSI, PALS and homegrown systems. The system will be brought up in 3 phases: University of Georgia (July 7), then 13 libraries, then the remainder. They have E4500 Sun Servers.

What they "don't have": report writer is a big issue because Endeavor uses Access or SQL. Circ expiration delay can’t be done like PALS. Previous patron on item record isn't the same. Can't discharge in backup. Billing report goes to a table of average prices for the item type or default. They don't use MARC Holdings Format for items - just holdings. They don't have ISO/ILL, but are partnering with Pennsylvania and Wisconsin to develop for Endeavor. The only way to get a union catalog including non-Voyager libraries is to create a separate database. They don't have universal borrowing and returning by any system library.

What they like: the customer base is growing, almost too fast for good support. It is a true multi-tiered client server system - they can have a database on one server, Acquisitions on another server or run it all on one. It is Windows95, 98, NT compliant and 32-bit [not all of it]. Can do simultaneous search of separate Voyager databases.

Migration: all patron's must have an address. Georgia expanded on work Tom Briggs did for Shippensburg. All bibs must have a 245. To convert reserves, items must have a title field. The 001 prefix had to be fixed, i.e. 00-, $00, 0cl7 had to be fixed to ocm. The 019 vs. the 001 as used by PALS won't match the number in the item; need to find all these and fix the item BIB ID's. The 090 field - it will use only the first one. Bibs without items can't migrate right because the location has to come from the item; you have to have everything barcoded. Free-format holdings - they had to write a program to identify them and input holdings manually; apparently it can't migrate to the field designated for it since they don't use MARC Holdings Format in items. Endeavor uses "Standard Information Format"; if you put your information in that format, then you can migrate. They have to do Acquisitions manually. Endeavor has a 100-day migration for everything. They do orientation with people who know the system. They get the data out, do a 4-day training overview of modules, then a test load. Endeavor uses call number, location, copy/vol. To determine MFHD then you check location (this was hardest), call number (if they took it from the bib, they got it in #h and #I), and odd-balls. Circ is easy to use, but hard to set up; policies are maintained by the system administrator per library. You get 5 days to review production loads so plan ahead what you are going to check. To migrate patrons, you must set up matrices RIGHT. Then you get functional training and finally advanced training. Acquisitions and Serials are complicated and they don't migrate.

Support has been good. There is no access to source code. Report writing is poor. Universal borrowing won't be available until 2001. There is a lot of cleanup that has to be done that is now a regular part of the work of Catalogers. It is not done by other staff, e.g. Circ staff can't fix items.

Gary will mail full page images of South Africa's presentation to each site.

Thankyou to MnSCU/PALS for setting up the facility and meals!

-------------%%%%%--------------

SAC Subcommittee on Subject Reference Structures

Sat. June 26, 9-12 AM

Thomas Mann

No matter how good the reference structure is or how wonderful the vendor's software, reference instruction needs to be given by librarians. If you show the subjects the way they are in print LCSH, the UF is not clearly not to be used and the fact there is a list makes some people feel if the search the general term, they'll get books on all the terms in the list and they won't. We need to teach users to search for words with the "tightest fit" - not the general word cuz it is only for books that only general works. We need to teach users how to use subject headings in their searches.

If you add general headings to bib records, how far back in more general headings do you go? There is no control or specified relationships when you get to such a general term that is would never have the cross referencing that would include the specific reference back to the specific thing you really wanted.

The Library of Congress has never completely filled out all the cross referencing structure possible for specific relations - often there is just a specific heading for a topic sitting there in the "red books".

There are a lot of specific terms that start with a word that are very narrow. You can't see the narrower terms that are adjacent in most online systems. [You can in PALS with BR SU cuz each one gets a line, but some systems list each instance of the subject - at LC you need to go through 242 screens of "business" to get to "business intelligence". Often there is no structure that would lead you to related terms, so that when you find a bib record for a book on the topic you see relevant (not related) subjects you could search.

Pre-coordinated subject strings present relations a user might never have thought of.

Free-floating headings are not usually in the "red books", but understanding them may be crucial to helping a user find what he/she wants.

We must not fall into the trap of believing that screen displays or vendors will remove the necessity of teaching.

"A catalog is a place where books get lost alphabetically"

Use of and formation of OPACS has changed little since the early 1980's. Users have developed adaptive behavior to bad systems. Few OPACS provide subject relationships. Users tend to get too much or nothing. They have navigational problems. Subject displays are only lists; they don't show relationships.

Enhancements you should look for:

• The catalog create order out of chaos with controlled vocabulary and cross references
• The OPAC should display the authority record before or at the beginning of the display of the list of items responding to their search - eliminates the problems of homonyms
• The amount of authority records varies by the type of library. There are 250,000 subjects in LCSH. 96% have a scope note. There are 2.45 references on most topical subjects. There are 1.8 see references and 1.1 see also's for geographic headings.

• Can enhance while cataloging
• Can buy things like TOC from BNA
• As content is added to the MARC record, the more successful the use by users

• We need a fundamental change
• Need to respond to queries based on user behavior: relevancy, search trees, etc., but not by quantity, but rather by quality - "best" terms (Karen Drabenstott, Marcia Bates)
• Should be able to do "find more like this"
• Natural language plus search aids

Enhancement of the user interface is necessary! Let's not keep replicating the card catalog-like displays

Kristin Gerhardt (Iowa State - was NOTIS, now Horizon)

She did research on Iowa State University's library OPAC authomated authority control on zero hits. What value are we adding by including authority records to the OPAC? They were able to measure a 10% drop in zero hits by adding authority records. Subjects made the most difference; while names and titles were less significant.

Zero hits result from:

• Error in the record (clean up your database!)
• We don't have the book
• Bad search

Analyzed the reports out of the OPAC against work being done in authority work and correcting bibs. The library added additional databases which also affected hit rate. Zero hit rates stayed about 22%; redirected searches went up then leveled off - i.e. people were following cross references and once they learned how to use authority records, failed searches leveled off.

Even if the library has no book for a heading, she feels it is better to show the user that we don't have anything but they DID a reasonable search and they can see related information.

Make sure users can search and browse LCSH files!

There have to be better display designs for OPAC's since they are available in other areas. We need to convince vendors of ILSs that we want it. We want standards followed.

Discussion:

Joe Matthews. Dumbing down to 1 or 2 subjects as often happens in "lower" level cataloging forces you often to pick more general headings and then you are out of the realm of the relational structure

-----%%%%%-----

Computer Files Discussion Group

11:30 AM-1:00PM

The Online … DG is trying to write definitions of "web page", "home page", etc.

Assigning subjects or form divisions or genre headings is often a difficult decision for electronic resources.

773 is being used in the PCC and EBSCO test [I tried contacting EBSCO but they said they wouldn't allow me access to the test file]

JSTOR titles and title changes aren't too hard because they have made them clear online

Infotrack holdings list isn't very accurate.

-----%%%%%-----

CORC panel

2-4PM

CORC - automated cataloging of electronic resources. Right now a lot of us are doing things locally in each of our libraries, but we aren't sharing. CORC is an effort at sharing. It has automated tools for: record creation, classification, subject headings, URL maintenance, authority control.

80 libraries are currently participating - ultimate test might be 200 libraries.

Karen Calhoun - Cornell

Dan Foley - Energy and Environmental Information Research Center

They have mostly text, numeric, and geo-spatial data. The focus is to get Dublin Core into web documents. Like CORC for searching, creating, and editing.

Duncan Irvine - Strathclyde

There are 800,000 websites and 1 million web documents. This overwhelming number causes resistance by catalogers to the idea of cataloging the internet. Reference librarians like pathfinders and reference service, not cataloging. Catalogers are under pressure to "get the books in and get them out".

The library's role is to provide selection and management of resources. The library OPAC should provide access to everything the library has or has access to. Their experience is that prior to OPAC's, users would go ask reference librarians for help; but, now, the user who doesn't find anything leaves the library.

The solid background of OCLC and the thoroughness of the project makes him confident in CORC. He likes the international shared environment. We can use agreed-upon standards, it is easy to use. Usually at least one Dewey number is close. They plan to catalog research being done on their campus, general internet resources, Scottish parliamentary documents and British documents. He hopes to use the process to strengthen the position of the library within the university system - it is a visible product and the library can be seen as a valuable partner in education.

OCLC eventually plant to synchronize the CORC records with MARC records in WorldCat.

-----%%%%%-----

CC:DA Task Force on Metadata

June 27, 8:30 AM-12:30 PM

Mary Larsgaard has all the signatures she needs for the pre-conference. It will be planned as a 2-day conference, $235.00 registration, OCLC is doing coffee. Looking at: changes in AACR2 and MARC21 to catalog web resources; practical training - go home and say this is what I need to know; "seriality"; needs for new rules and tags; alternative approaches - CORC, etc.; ISO standards; mixed groups - traditional library, art or EAD; "vision". Full plan must be in by Oct.

This committee is using from its charge: #1-3 to investigate #4, to make recommendation #5.

Group #1. Interoperability is going to be critical in the future - cross-mapping not only for a bib record, but also for search engines. Users define who they are, we don't. One user is both novice and advanced. What we learn from CORC could be very interesting. Eric Childress commented on friction between MARC and Dublin Core and how to crosswalk; is DC okay or will MARC end up being needed for people to find resources; people seem to be trying to force MARC into Dublin Core.

How do we want to proceed when MARC records [or should they be XML records??] can be created using multiple standards with AACR2 being only one of them? Could use FDG-, AACR2. Can we use it? Is it good enough?

A big problem: the exchange of information has been hindered by incompatible hardware, software, protocols.

Each description community shares structural, semantic and syntactic standards: libraries did AARC and MARC; other communities have developed other systems. W3C is trying to develop standards to cross communities' standards. Functional requirements will evolve! Need to define commonalities.

XML is a language, a mechanism to define tags and the structural relationships between them in documents; it's extensible. An XML record says: I have these tags, how they are formed and used in this record. XML has broad industry acceptance. It is a subset of SGM. It will be in Microsoft Office 2000 Word and Excel. XML is a lightweight SGML. It is the new ASCII. XML can include a document type definition (DTD) - i.e. says what's valid vs. not valid; it is not required to include DTD's as long as the record has "wellformedness" (i.e. open and close marks).

W3C they are working on XML schemes to define certain DTD's, e.g. in a Dewey number the decimal is a "float" (like Excel) and it must be a certain type of number.

XML is not a display thing or a print thing. DOM = document object model; with DOM you can include HTML display. CSS = cascading style sheets; a different mechanism than HTML for display. XSL = extensible style sheets; goal of this new styl sheet being developed is to handle display and print.

XML data transmission. Where would I put the style sheet information? Could be within the document, or could link to a http header screen, or could get style sheets from a trusted third party. So, I could have an XML document and a) if I can see, I can go to a server that says how to display XML on a visual screen, or b) if I am blind, I can get it in Braille.

Sometimes syntax isn't enough: just because you agreed on a syntax (MARC), if you don't get certain information (LDR, 006, 008) you couldn't do anything with it.

If each community can create its own structural, semantic and syntactic standards, how do we put it all together? RDF is a data model. It is like a box, a leggo. Within it, each community can define its own DTD, but they must be able to remain separate and to a standard. The outside of the leggo has the abilty to match up with something else on another leggo. The RDF says I am a leggo - and that's it. The schema used within it can be determined by a community which can delare its vocabulary - stored at a website.

An RDF statement can only be written one way. URI (this web address) has value of (properties of) this resource or string value.

This resource (document) has a title defined by Dublin Core (property is a type of resource) with the value of title: ALA presentation

This resource (document) has a creator defined by Dublin Core with the value of Eric Miller [this could be another resource URI)

Namespace = place to go to find vocabularies, standards, schemes; it is a way of defining context

RDF has its own basic set of vocabulary you can use; when you want to use something else, you declare it in the top part - you could make up your own - we in the library community need to decide what we want to follow

XML uses Unicode therefore can handle multiple fonts, languages, etc., but because it is Unicode, capitalization is important.

RDF uses "link authorities", i.e. we can identify Eric Miller from other Eric's because we can link to a URI (value that is a resource) with Eric Miller's email. RDF is anything that can be uniquely identified.

There are a growing number of tools we can use with XML and RDF. It will work with all browsers.

There are still issues to address: work vs. manifestation; multiple languages; aggregations - collection vs. item and how to link; how we set up the library infrastructure to deal with different link authorities (e.g. AAT online, Getty, etc.)

What do we have to offer in the world environment? LIBRARY LEGGOS

-----%%%%%----

CORC participants breakfast

June 28, 7 AM

So far there are 45 academic, 3 state, 17 federal, 2 school, 8 non-US, 5 public and 9 OCLC networks participating. New features include constant data and update authority files although it is still not concurrent.

New in MARC: can import, authority is available, full page editing, better fixed field editing

New in Dublin Core: better export, refined mappings to MARC, MARC subfields visible, continue to add qualified elements

Pathfinder: can put image on top of page, more display options, objects links to CORC records

Immediate plans: WorldCat synchronization, new pathfinder editor, TEI header loading and export, Unicode (authorities are unicode, but records aren't yet), language identification, URL checking i.e. reporting, phrase extraction, simplified subject headings - i.e. not pre-coordinated.

Future: authority records, GEM, IMS, MeSH vocabularies, distributed CORC, goal is in production next summer

Suggestions from CORC meeting: metadata guidelines; URL maintenance; auto LCC, NLM, etc.; additional thesauri; automatic PURL's; patron access; searching changes; WorldCat NAF; full page editing; easier pathfinder editing; constant data template; diacritics; images in pathfinder

WorldCat synchronization: looking at something like a place to enter my WorldCat authorization and then I could "push" the record from CORC to WorldCat

Looking at Z39.50 access.

Faceted Application of Subject Terminology (FAST) - simplified headings, i.e. would need to be post-coordinated, but it fits better into Dublin Core because some elements are for information we put into LCSH subject strings, especially #y and #z and #v

Will the patron be better served, should be the question. We sit here and say users don't know how to search strings. We say it is too hard to train inputters. Users do lots of subject keyword searches in systems. Is this a problem of system design or subject strings ???

(left to go to SAC meeting)

-------------%%%%%------------

June 27 PM

I stopped at the OCLC booth and discussed CORC. I also reviewed a new web interface that would be useful for small libraries in North Dakota to do copy cataloging. There are only a few options to change on the records. They would need to be downloaded. They remain "saved" until downloaded, so the ODIN Office might be able to download on a weekly basis using each library's password.

I visited with Endeavor. I asked if they are looking at XML. They guy I got said, yes it is being discussed. I asked him to show me cataloging. He had no clue.

-----%%%%%-----

SAC Subcommittee on Metadata and Subject Analysis

Mon. June 28, 8:30AM

Program for 2001 on metadata and subject analysis will be planned by this group, beginning at midwinter 2000.

We are co-sponsor of pre-conference on metadata.

Should cross references in thesauri have a role in the web?

• couldn't they be built into search engines?
• Most browsers often suggest topic areas - or - "these are other words you might want", so this isn't an alien thought
• Most browsers go up or down - with thesauri, you could go sideways; even with classification schemes you would tend to just go up and down
• EPA is developing its own thesaurus with narrower, broader, special
• Relational terms are being included to enhance usability

Metathesaurus section

• the idea is to identify and use or crosslink, etc. existing thesauri - don't start from scratch - make different vocabularies work together; "weave vocabularies into a tapestry" of appropriate terms
• we need to be clear on what our definitions are for: harmonization, metathesaurus, mapping or linking, cross-walks, controlled vocabulary
• metathesaurus sort of implies some big huge list
• harmonization sort of implies smooshing together
• linking and mapping is a structural method of crosswalking; if we map, we leave it open for people to choose words that may actually lead them to different types of published works, because "cancer" may get more popular works than "neoplasms" will
• one is automatic
• harmonization
• will probably get big lists
• one is optional
• linked
• doesn't always crosswalk well
• IFLA project, Multiple Subject Entry (MACS) is trying to harmonize subjects in bibliographic records in 3 languages using 3 subject thesauri (LCSH in English, Rameau in French, and Schlagwortnormedatei (SWD) in German)

Do we want to encourage the untrained "cataloger" to assign top level classification data?

• Mechanisms would assist people in assigning top-level [i.e. "appropriate"] classifications - do we want to do this? Will it be too general or does it matter?
• Class scheme abridgements could be made available

Add to III System

• Online tool non-catalogers could use to suggest an appropriate subject term or classification number

Suggest using numbers and subnumbers for the sections to pagination doesn't matter. A table of contents should be included. Include preference for "narrower, broader, related" relationships. Plan to have this ready by IFLA Aug. meeting. We should be able to do an email vote if changes are minor; then send to SAC for approval and finally CCS Exec for approval. OCLC is very interested in what we report.

Shannon Hoffman moved that it be okay to share our working draft with OCLC, IFLA, and LC, making sure it is clear it is not an official ALCTS position.

What's next for our committee?

• Use of related terms
• "harmonizing" thesauri
• any relation to classification & metadata committee?
• When, how, do you break up a pre-coordinated subject string?
• How would people reach a thesaurus - that is a system design issue, or perhaps part of the Subject Reference Structure Committee

Next meeting

• Program planning
• Deconstructing:
• what is OCLC doing and what does their research show; should we make any recommendations?
• Should we look at something beside Dublin Core?
• Should we look at other thesauri and if so, what that thesauri are there?
• We could bring in someone to discuss several thesauri
• Decided to:
• Overview of a number of different metadata schemes, similarities and differences (Bob Pillow)
• Program planning
• Report of FAST
• Concluding report

OCLC representatives, Diane Vizine-Goetz and Eric Childress

• Faceted Application of Subject Terminology (FAST) is part of a test in CORC to de-coordinate subject strings and to be put in Dublin Core
• Subject = subject, (probably include #x)
• Type = #v (nature or genre of the content of the resource)
• Coverage = #y or #z (spatial location (a place name or geographic co-ordinates) temporal period (date or date range or period) or jurisdiction (named admin agency)
• Format #v? (media type or dimensions)
• Geographic names could link to Geographic Area Codes list
• Diane will forward information from OCLC regarding FAST to SAC-L and in CORC-L archives
• Need to be able to cross-walk between MARC and Dublin Core
• An idea of CORC is that reference librarians could create a CORC draft record and put in terms that the cataloger could use to create LCSH

Meeting time next ALA will be Friday 6-8PM

-----%%%%%-----

Committee to Study Serials Cataloging

Mon. June 28, 2-4 PM (able to attend only 2-3PM)

CC:DA

• Barbara's report to CC:DA will be on CC:DA's web page; JSC reports are also referred to there
• AACR2 (JSC & CC:DA) should be brought into harmony with international standards (ISBD(ER)) -- updates will probably be available free on a web page; final report should be up by July 31
• 0.24 will soon be on JSC's page
• the definition of monographic series will be drawn from discussions
• all remote electronic resources will be considered published
• gmd: decided not to say "electronic tape cassette"; use "computer", not "electronic"
• CC:DA will have a committee to respond to Jean's proposal
• 0.24 and Martha Yee's paper addresses problems:
• problem applying rule to things with multiple characteristics
• work in multiple manifestations (multiple versions)
• option 2 was chosen: treat content and carrier as equals, follow the Delsey model; a new text of 0.24 will be put out quickly, whereas the rewrite of AACR will follow later
• whenever a specific rule varies from a general, follow the specific
• options to deal with "major" or "minor" variations, approved option 5:
• new record is work or expression
• not a new record (except there will be an option for manifestation variation) is manifestation or item
• decided:
• create new record for expresion
• create new record for different incarnation - what is "incarnation"? - a new word to replace expression for serials
• all changes to an integrating entity should be on one record, latest entry

News from Library of Congress

• Aug. 16 their new ILS comes up
• Serials will still be done on OCLC because of CONSER membership
• They can't access CONSER through Endeavor; they can only get to it on the public access side through OCLC SiteSearch
• Pre-MARC, serloc, CONSER will all be in one file in the new online
• Conversion needs to be done for serials check in; estimate is it will take 5-7 years
• Pinyan - the geographical locations are now Pinyan; headings still need to be updated
• PCC - the Serials train-the-trainer manuals are now available
• Faxon has joined CONSER
• CDS is producing a new CONSER Editing Guide and new Cataloging Guide
• Publication patters for 853, 863 will be put in OCLC records by CONSER libraries, but may prefer to post on a web site somewhere
• Watch for the next LC Newsline
• Beginning ISSN manual revision and harmonization with JSC report - NSDP still prefers separate ISSN

MARBI Report

• DP114 - seriality and MARC21
• Identify integrating records in MARC records - eliminated proposal of a new bib level code, in favor of changing "m" and "s" or making a newType code
• For the 260, making 260 repeatable and using an indicator to indicate old or current
• Waiting for JSC decision on title changes re: the 008/34 for successive or latest entry
• 99-08 repeatable URL added to 555, 583, 843, 76x-78x
• 99-09 approved 852 #k #m repeatable
• DP116 "bound with" is about a physical entity so only 87X will be used to link bound withs; no repeatable 004 will be used

Jean Hirons began her report (important!) but I had to go to EPSCoR

EPSCoR

Mon. June 28, 3-5PM

Elaine Albright, Gloriana St. Clair , Juana Young, Janet Parsh, Bob Neville, Warren Murphy, Jessica, Shelby Harken

Arkansas and Nebraska looked at "recombinant DNA technology". Arkansas made a list of subject headings and wearched WorldCat (OCLC). They identified those that would not be a copyright problem. They also search NTIS, GPO, BIP and Journal Citation Reports. They check OCLC for which titles are held by Universities with doctorate, graduate, and undergraduate programs based on the Gourman Reports.

Wyoming reported that they tried a similar process with "mining" but found extensive overlap with other topics. They searched EI compenix, CSGS, WorldCat, GeoRef, EPA, US Bureau of Mines, state publications. The Office of Surface Mines and Reclamation in Pittsburgh is digiting mine maps. USGS plans to scan some early USGS documents.

Michael Lesk was denied the ability to participate in the project as proposed - he received a letter from Joe Danek (EPSCoR) indicating the project should focus on networking and supercomputing and faculty input was needed. Librarians did not receive such a letter. Despite the letter "we" (Elaine and Gloriana) feel we should still do the pilot project. As long as Michael is around they might be able to get $1-1.5 million of non-EPSCoR monies. Elaine would like to continue at least as a pilot. Attendees had lots of reservations about the project as a whole, but not so much about a pilot. A pilot project would have to be very successful and be done well and quickly to show value. If it appears to be of little value, then that would be the end of the EPSCoR project.

NSF may care about the idea enough to fund "somebody" else if EPSCoR doesn't do it. Elaine said that EPSCoR libraries have the opportunity to do a project that extends research of university scientists; it gives us the opportunity to report back to "those that be" positively. If two or three or four libraries want to do the pilot we should be able to get money; monies should go to those doing the pilot. The other libraries should write letters of support. Libraries are a representative pieces of the institution that says it's a member of the EPSCoR institutional coalition. We (Elaine) need a broad enough topic to include 5-6 people but not so broad as to be so big we'd never get it done. We need to pick a topic and see what are the strengths. Elaine would send out a letter saying for example what two topics are and libraries would respond saying if they have materials to support digitizing in that area.

We are giving up staff time and our print copy of the book. Gloriana felt $25 per book was a good estimate, but attendees scoffed at that cost saying evaluation, mounting the web, etc. would cost much more. Work could be done at each library or mailed to a library doing the digitizing. We would ask for money for scanners. She doesn't think we can ask for money to buy copyright. We have to make sure we have material that scientists want, i.e. current materials. Grey literature could be used, or materials that are hard to get in the first place.

The project with 3-4 libraries would need to do 10,000 books in 3 years. Costs will included $50 per book, 50% overhead, acquisition of 10,000 books, rebinding books, getting permission to digitize, server costs. We need to prove such a project is useful and we learn as we go what it costs.

Elaine wants to send out a letter to EPSCOR libraries asking them if they can be a pilot, if they have a collection to contribute, if they're willing to give up their books, asking if the volume of books in your collection will be enough - need to make a bibliography of 10,000 items. Then librarians and faculty have to identify 10,000 GOOD books, then staff has to scan in the books provided you can get copyright permission. Maybe it could be done in 2 years. Doing the metadata includes cataloging and creating bibliographic records and using XML, setting up the web site and page design with hypertexted contents and links, and reviewing the intellectual properties (copyright). Carnegie Mellon is willing to serve as a consultant.

We have to show EPSCoR interest on our campuses that we are willing to do this and "take it out of our hide."

If there aren't 4-5 libraries willing to participate in a pilot, then we have to wonder if this is even a possible project. Janet felt the letter will be used by some EPSCoR libraries to mean they should not participate and they will use it to get out of the project. Each library needs to be to say how may staff and how many books and in what time frame they are willing to give up to the project. Then we can decide the scope or definition of the project. NSF has said it wants to see a pilot; EPSCoR has said it doesn't. So it is EPSCoR we are trying to convince we want money to do this because librarians think it is worth it. [If users don't think so, it isn't worth it]

Next meeting by compressed video.

Questions or comments: Shelby E. Harken

Back to ALA, etc.

 7/8/99