1. APPROVAL OF MINUTES FROM 2-19 AND 2-25 CONFERENCE CALLS. 
The minutes for February 19 have been approved.  The minutes from February 25 need 
to be revised regarding the 856 subfield order and options for display and will 
be distributed for the next meeting. 

 2. NEWS AND ANNOUNCEMENTS 
Alan announced that he received an updated list of current Heads of Cataloging.  A 
copy of our survey has already been sent to each of them. 

Re: Summary of NISO standard 
Alan offered to mount his summary of the NISO standard on the web site soon.  It 
can be amended in the future if necessary. 

 3.  856 SUBFIELD ORDER 
French announced that all have agreed with CDL's recommended display order and the 
minutes of Feb. 25 have been changed to reflect this. 

 4.  MELVYL RECORD PROCESSING AND DISPLAY. 
This topic was postponed until the next meeting because of Doherty's absence. 
Instead, Graham reported on processing aspects related to UCSD record 
distribution.  She summarized the contents of her email report (inserted below): 

"Here's an informal update from UCSD that contains lots of hard questions. 

I met with the UCSD Systems staff to discuss the resources we would need to 
serve as the Central Cataloging Agency (CCA) for the CDL.  Our meeting was 
held before Rebecca Doherty reported that CDL would likely be willing to handle 
record distribution, which seems like a better alternative than having UCSD 
send records directly for the campuses, as explained below. 

Currently UCSD ftp's weekly updates to CDL.  The data is compiled using 
Innopac's OUTRECS file, where all the cataloging transactions for the week are 
collected.  (This is a real-time file, with the latest version of designated 
records; there are no duplicates).  We then subsearch the file, eliminating 
records we do not wish to send (i.e., suppressed records) and divide into two 
files: periodicals and monographs.  We ftp these two files, along with an 
email message with the file name and record count.  We leave our files on the ftp 
server until we receive a message from CDL/T that the files are loaded on 
MELVYL and we can delete them (about two weeks). 

The easiest method for us would be define location code(s) for CDL and include 
the CDL records in our regular weekly updates.  CDL/T would display this 
code(s) as "CDL Internet" instead of "UCSD Internet."  We need to verify 
whether CDL/T can handle that or whether we should send the CDL records 
separately from our regular weekly updates.  If we need to send them 
separately, what, if any, customization do we need to do?  If we need a 
different profile for these records, this will incur a charge of $2500 (plus 
staff time).  At one time Rebecca mentioned that these records would be 
treated as a separate campus.  It's unclear to me if that's just a different label 
(CDL Internet) that could be derived from records in UCSD's regular output or 
whether you need some other designation (e.g., an 049 code).   Please advise. 

Note that the CDL records are separate bibliographic records from our UCSD 
records.  We would code them with a special RCTRL CODE ("j") which means they 
display in MELVYL but not in ROGER (our local catalog).  While the data 
will be similar to our local records, the records are independent and local 
updates will not trigger updates to the CDL records.  (That is, if we change our 
PRINT holdings, that won't generate a replacement CDL record).  Since our local 
record will also go to MELVYL, we hope that CDL/T can give weight to the CDL 
record so that it takes precedence over campus records when selecting the base 
record. 

I was quite relieved to learn that CDL/T can do the campus distribution for 
several reasons: UCSD has a shortage of Boolean files so creating separate 
file(s) for each campus would probably require that we purchase more files. 
(They are sold in lots of 40 and cost $2500 per lot).  We also have a shortage 
of space on the server so we would not be able to retain the files for 3 
months as CDL/T can.  If we had to store files for any length of time, we'd have 
to charge for the space.  CDL/T seems prepared to handle the local data (i.e., 
manipulating it for MELVYL display and stripping it before outputting to 
campuses)." 
 

Graham's report was followed by a discussion of OCLC holdings symbols in CDL 
records. Because we have made the decision that the CCA should utilize OCLC 
records when cataloging CDL records, we need to address the issue of OCLC 
requiring holdings symbols for all records.  These symbols function primarily for 
Inter-Library Loan.  They also are the mechanism OCLC uses to charge for use of 
records.  Attaching holdings symbols to records for electronic resources is not 
desirable because most cannot be lent via ILL.  CDL licensing restricts access to 
UC users, so use or distribution outside UC is not allowed.  It would be even more 
confusing if the CDL holdings symbol were added to a print record which has been 
used to represent the electronic access.  CDL would never actually own the print 
material. UCSD has been discussing this matter with OCLC, and we are awaiting a 
decision. 

One possibility would be to add only the CDL holdings symbol and list the CDL as a 
non-provider for ILL.  The holdings byte question applies to locally-cataloged 
Internet materials as well as CDL titles, but the non-supplier option is 
inappropriate for locally-cataloged titles.  We agreed to refer the question to 
HOTS. 

For the purposes of the report, the group agreed we should use wording 
similar to  "Pending HOTS recommendation regarding the OCLC  holdings symbol for 
records, TFER recommends that original and copy cataloging be conducted in 
OCLC, with the CCA meeting contractual obligations as determined." 

ACTION: Alan will put this on the agenda for the April HOTS meeting. 

 5. RECORD DISTRIBUTION MECHANISM 
There are further questions to answer regarding the type of customization required 
prior to distributing records from the CCA.  In general, individual customization 
of data elements cannot be offered.  We still need to define what the CDL data set 
will be in terms of the specific fields that may need to be stripped out of 
distributed records (mainly local, Innopac fields?) and CDL fields to include 
(e.g., call numbers).  One type of customization that will be supported is sorting 
the records into campus sets containing the records for the CDL titles to which 
each campus has access. It was suggested that we recommend staging a pilot project to test record distribution and local holdings. UCD has offered to be a test site for local receipt and loading. 

 6. DISCUSSION OF AREAS/ISSUES WE WANT TO HIGHLIGHT IN OUR REPORT AS IMPORTANT FOR FUTURE ACCESS 

We agreed that we will want to include recommendations for extending the work of 
this task force and other issues that need further study.  Among them are: 
        Government documents 
        OAC/EAD Collection Level Records 
        Digitally archived content 
        Maps and imagery 
        Aggregator databases 
Graham has provided us with an overview of the issues associated with aggregator 
databases (quoted from email, sent to TFER 2 and all former participants in 
Full-Text Project): 

"This message is intended to stimulate discussion about aggregator databases. 

Aggregator databases are large databases that aggregate, or bring together, 
the full-text of journal articles.  In most (all?) cases the journals have print 
equivalents.  The aggregators fall into two broad categories, those that 
aggregate the publications of a single publisher (or occasionally a small 
group of publishers).  Nearly all of CDL ejournals are aggregates licensed through 
a publisher (e.g., Academic Press, American Mathematical Society, Wiley, etc.) 
The second type are subject-based aggregators.  These aggregations include the 
full-text of articles published by a vast array of publishers and made 
available through a third party (e.g., ABI/Inform, MAGS). 

The publisher-based aggregators are somewhat more stable than subject-based 
aggregators, although coverage is constantly changing.  The publishers vary in 
their levels of communication with libraries, but they are beginning to 
provide "what's new" pages on their web sites.  Most, but not all, publisher-based 
aggregator databases retain the concept of "issues" so the cataloger can 
discover what time-span is included. 

Subject-based aggregators are more difficult to describe.  Content of many of 
these titles no longer exist as discrete issues and be accessed only by 
doing a subject search across the databases' content.  (In Melvyl one can 
collocate articles through "xjo" searches but few users are aware of the more 
esoteric search keys).  There is no chief source to provide bibliographic data so 
catalogers cannot prepare a bibliographic description according to the 
traditional library model for descriptive cataloging. 

CONSER has formed a task group to gather data and make recommendations about 
what CONSER could do to provide bibliographic control for these materials. 
The group includes Ruth Haas (Harvard), John Reimer (Georgia), Jeanne Baker 
(Maryland), Karen Calhoun (Cornell) and Jean Hirons (LC).  Some of you may 
have attended a discussion on this at Midwinter, where these individuals 
presented a variety of opinions.  (Already distributed were Jeanne Baker's list of 
reasons to make separate records and John Reimer's survey results).  Options 
include 
1) CONSER's single-record approach; 
2) separate records for each journal title; 
3) adding 510 fields to the bibliographic records for the journals 
4) adding holdings to the print records. 
Each of these options has serious problems, such as instability, distribution, 
difficulty of description, record length, and the desire to include web links to 
the full-text.  CONSER seems to be focusing on subject-based aggregators at this 
time. 

The current issue before us is what to do with the updates to the full-text 
project we did several years ago.  At the time we created records for 
ejournals in MAGS, ABI/Inform, and COMP, we did not yet have Web-based OPACs.  We decided to simply add a holdings segment to the print record for each title 
indicating availability and coverage of each title in the database.  The original 
work was done cooperatively with 7 of the 9 campuses taking responsibility for a 
group of titles.  Some campuses went on to add the data to all their records so it 
would be available in the local catalog as well as MELVYL. 

CDL/T has now sent us a list of changes to the data (new titles, changes in 
coverage, addition of full-text where there were formerly just abstracts).  We 
need to figure out whether we want to add the updating to the CCA 
responsibilities, do another cooperative project, or take another approach. 

Sally Tseng of Irvine has suggested that we include the 856 for the MELVYL 
database in the record for the print.  As it now stands, the MELVYL PE user 
must log into a different database (a MELVYL A&I database) to get to the 
full-text.  A user of a local OPAC must log into MELVYL.  But if the records 
included 856 fields, the user of the Web-based OPAC could click and get into 
the database directly.  (Unfortunately they couldn't go to the specific title 
but at least they'd get the MELVYL catalog home page with the correct database 
already selected). 

If we were to take this approach, we'd need to decide if we wanted to also add 
007, 530, etc., so they would conform to the same standards as the CDL 
records, or if we would just add an 856.  The location symbols were programmed 
differently so we'd need to work with Rebecca to discuss the programming 
implications if we wanted to take a new approach." 
 

French questioned whether aggregator databases are always meant to mean full text 
databases or whether conglomerate web sites (such as Cambridge Scientific 
Abstracts) fall into this category.  CONSER's report seems to focus on full text 
specifically. 

The Melvyl fulltext titles in MAGS, ABI, and COMP are another form of aggregated 
databases which we dealt with 4 years ago. by creating Melvyl holdings segments 
in PE.  This project is due to be updated and Graham questioned whether we should 
recommend that it be included in CDL record distribution from the CCA.  There has 
also been a suggestion that 856s be added to the title records to connect users to 
the Melvyl database which has the full text.  We decided to refer this question to 
HOTS and Alan will put it on the next HOTS agenda. 

 7.  CDL DIRECTORY LINKING 
This topic was deferred until the next meeting due to Hixson's absence. 
 

The next conference call will be on Monday, March 15 from 2-4 PM.