Notes from the IO Cataloging Congress, Aug

Notes from the IO Cataloging Congress, Aug. 24, 2000

Radio & TV (IUB) Room 251 10:00 a.m. to 3:30 p.m.

Approval of minutes: The minutes of the July 21, 2000 meeting were approved after being amended to include a statement thanking Chris Long for taking minutes.

Agency Status Reports: Marty Joachim reported that several sets of microfilm records have been partly loaded in BB and put through vended authority control. Some cleanup will be needed. They include both microfilm and microfiche.

Sirsi implementation timeline update: Gary Charbonneau discussed the upcoming cataloging moratorium.

Sept. 15, 2000: Moratorium on all UITS bib and auth batch loads

(Excludes MARCIVE records for Indianapolis Law

Library, NW, SE and Peter Ward LC resource file

updates; these will continue until Oct. 20th).

Oct. 20, 2000

Effective 5 p.m.: The following activities in NOTIS may no longer be performed:

Creation of new bibliographic records
Editing/deletion of existing bibliographic records
Creation of new authority records
Editing/deletion of existing authority records
Creation of new copy holding statements
Editing/deletion of existing copy holdings statements
Creation of new MHLD (MARC holding) records
Editing/deletion of existing MHLD records
Creation of new linked item records
Editing/deletion of existing linked item records
Creation of new unlinked item records**
Editing/deletion of existing unlinked item records

**Oct. 23-Dec. 22, 2000: May resume the creation of new unlinked item records solely for the purpose of circulation (including rush cataloged materials during the moratorium)

Dec. 22, 2000 (effective 5 p.m.): No longer may create new unlinked item records

Dec. 26, 2000: May be able to bring up Cataloging

Jan. 2, 2001: May resume normal cataloging activities in Unicorn. We may give the public a date of one week later.

Cataloging data will be extracted after 5:00 p.m. on Oct. 20. On the following weekend, we must not do anything with unlinked item records because they will also be extracted for processing. They will have to be converted to bib records. On Dec. 22, there will be a second extract of unlinked item records created between Oct. 23-Dec. 21. Do not create new ones unnecessarily.

Horizon will essentially follow the same schedule, except for MARC holdings records. In Horizon these are tied to predictive check-in, which NOTIS doesn’t use.

Acquisitions processes: There will be an extract of acquisitions data on Dec. 1. We can pay for materials until Nov. 14; check-in can continue until Dec. 1. We can’t create new orders or do anything that changes the copy holdings record.

Security in NOTIS: The bib file will be closed for record creation and editing. Do not do anything to item records after Oct. 20 because your changes will not be reflected in the load.

Horizon users could use old McBee cards for circulation. Circulation will go on. All things that are charged will go over to Unicorn on Dec. 22.

Authorities:

Spencer contrasted authority records in NOTIS and Unicorn. The main difference in the records themselves is in the headings use codes; “c” will not be used in Unicorn. It was used in NOTIS to turn on references in the OPAC. The fixed field labels are slightly different, but the data comes across. The 690 fields in authority records will be protected and retained when overlaying from OCLC. The 690s that showed we added local references will stay, but the references themselves will no longer be easily identified, since the NOTIS repeat number drops off (410/1, 400/2, etc.). All authority records in NOTIS will be brought across, even one-liners. The BNA one-liners have already been removed, but if any editing took place on one-liners, they were not removed and will be retained.

Searching for authorities in NOTIS was contrasted with searching for them in Unicorn. In NOTIS, authority records, references and bibliographic records all file together. In Unicorn, headings and entries are divorced. The authority index is made up of 1xx and 4xx entries. The 5xx entries don’t display in the technical services mode, but they do display in the OPAC. The “A” at the left side of a heading in Unicorn designates an established heading; the rest of the headings are see references. The system displays diacritics and the ampersand, but you must leave out punctuation and the ampersand when you search.

The control tab on the left side of the record stores the record number, which is generally the 010 field. Those having no 010 will have a system-generated number starting with “xx.”

The “created by” and “modified by” areas show who last modified the record. The “date created” is the load date; other dates in the NOTIS record won’t be retained. The default date of “never” could possibly be used in acquisitions to filter out headings for authority work.

The 690 field might still be a good place to store various kinds of local information.

There are three main authority indexes in Unicorn:

(1)LCNAMES (100,110,111,151,130)

(2)MESH (150)—coded for MESH in the fixed field

(3)LCSUBJECT (150) Not having 151 in the subject index could be a problem on indexing; 151 headings can also be 110. It might be good to have 151 headings in both the name index and the subject index.

A heading can’t be posted to a different index based on the presence or absence of a subdivision.

All 4xx headings are indexed, but that doesn’t mean they all display in the OPAC; Unicorn works just as NOTIS works in that regard.

Entries are normalized, even though the punctuation does display. In searching, Unicorn doesn’t require a period and a space before a title following an entry, as NOTIS does, or the hyphen tacked on to the subdivision of a subject.

You can browse backward and forward and go to the beginning of a display, but you can’t just jump to a certain number in the display, as you can in NOTIS. There are fuzzy matches on implicit searches, and a right hand truncation on the browse.

“See” references come from thesauri. The thesauri are authors, titles, and subjects and are built from authority records.

In verification, “unauthorized” sometimes comes up. A heading is either authorized or unauthorized, and they interfile, e.g., |b and |d normalize. In NOTIS they don’t file in the same place. Spacing matters in verification, e.g., A.C. McClurg will file apart from A. C. McClurg. Workflows is also picky about diacritics and capitalization. But the indicators can be wrong and still pass verification. An asterisk beside the heading in the validate list means it’s an authorized heading. “Validate” is something like CLARR, but it doesn’t grab a new heading and bring it in. There is no fuzzy match on authorities.

The “?” subfield doesn’t display, but it contains the word “unauthorized” if that is the case. It helps in “reporting functions.” Don’t tamper with the “?’ subfield itself.

The equal sign (=) subfield shows only in reports. It’s an internal pointer to the record where the authorized heading is located.

If there are multiple authority records for the same thing, it points to the first one.

You can’t just look at the thesaurus; it doesn’t exist alone. When you are browsing, you see what is generated by your search.

Use the “edit” font; it has all the diacritics as they really appear, and the diacritic comes first, followed by the letter. The “display” font is simpler, and they are composite characters, but it displays the diacritics incorrectly.

An asterisk (*) before a heading means it’s an authorized form. A question mark (?) means it is a cross reference. From the index, there is no good way to tell if the authority is from LC or is a local authority record. But it shows in the “display” font. It is hoped that the system will have the “edit” font in all aspects (OPAC, cataloging, acquisitions). In some areas, the display font is automatic, but mostly we will be using the edit font.

There is no pull down menu with diacritics, but you can cut and paste from the authority file.

When it creates an authority record, it doesn’t make a 670 field as the NAR function of CLARR does, but Sirsi is supposed to develop a wizard that creates see references and a 670.

You can generate a new heading from your record and set it to “provisional,” so that you can go back and validate it later.

Indexing:

The hyphen is significant punctuation; you must use it or you will be placed in the wrong place in the index.

The index display of series shows series numbers on the screen, but they are not in numerical order.

Bib records are stored as “last in, first out,” meaning that the latest record to be added in a grouping will show at the top of the index.

There is no “search history” feature.

There is keyword searching, both in the staff mode and the public mode, so it is possible to find a particular number in a series.

Uniform titles: The authority record must have a 1xx and the title following it in order to match. This works on 6xx and 7xx uniform titles. It doesn’t match on a 100 field plus a 240 field, just like NOTIS.

Authority records: It will be easier to do NACO records in OCLC and then just bring them in to Unicorn rather than working in Unicorn. Currently there is no way to upload an authority record from Unicorn to OCLC. You can create a provisional authority record. The best time to do the authority work is up front.

When you delete an authority record and the bib record is left, then the heading on the bib record becomes unauthorized.

If you change a heading on an authority record, it will go back to the related bib record and change it and all related bibs overnight. There must be an exact match, though. It won’t change headings followed (for example) by a |t and a title. You will still have to change those with subdivision |x etc.

There is a global change feature, but Spencer said it “isn’t pretty,” and he hasn’t tested it yet; it isn't believed to be as powerful, or useful, as NOTIS' glob/glch commands.

The “validate” button will make it link up to the authority record, but we do not see this link. The authority record number is stored in the bib record.

The newest record goes to the top of the sort unless you request a different sort. In the OPAC it defaults to an alphabetical sort by title.

There is complete keyword access to every part of the bib record, but you can’t go to the subfield level.

Peter Ward Files: We can get these and put them in a separate database and then bring authority records into Unicorn. It doesn’t automatically capture them as the BAM feature of CLARR does. We will have this file in January, but Mechael isn’t sure if it will continue.

Unicorn authority reports: Gary had a handout; the first page showed what was asked for and what it did. Page 2 showed which headings on the record were unauthorized. Page 3 was an example of a sort generated by cataloging date (greater than Aug. 15) and by Bloomington Main Library. You can generate reports as many times as you want; it’s not just a “new headings” list as in NOTIS, and you can break it down by library. You can adapt the request to fields under authority control, and you can use record ID numbers to run it. Page 4 showed an example of a record run from the report. We could use the date cataloged and the modified date for authority control. The default on the catalog record is “never,” and it would be best if Acquisitions staff left it that way.

Who can run these reports? UITS will not run them routinely unless we ask them to; they will not be printed from a central location. Circulation reports could be run by the circulation desk supervisor. They may not need to print the reports; they might just view them or edit them and print only what they want. Decisions must be made as to who will set up the reports, how useful will they be, how should they be formatted, and how often should they be run. They will need to be run at night because the report is working against the real database.

Authority Control (Mechael Charbonneau) It has not been decided how we are going to do authority control in Unicorn.

There are many blind references when you restrict your search to one location. There is no way to turn off the display of references when there are no corresponding bib records. Keyword searching is better if you’ve done good authority control.

We could consider vended authority. We are now down to three people doing regular database maintenance in NOTIS, and we are already using some vended authority control. Mechael called for volunteers for a group to look at vended authority control. OCLC has BNA now. What are the costs and advantages? Marty Joachim, Joe Harmon, Sylvia Turchyn, Linda Kelsey, and Ria Collee (chair) volunteered. A charge will be written, and the report should be finished by February.

Documentation (Mechael). The old IOCC documentation which was written over the years for NOTIS is on the Web, and a group needs to go through it to see what pieces of documentation are still valid, which are not valid, and which need to be edited or re-written. Scott Opasik and James Castrataro volunteered to do the initial review with a November deadline, and then other groups would take over.

Cleanup: On October 2, the reports may be re-run. However, if the data warehouse isn’t fixed by then, the reports won’t show the fixes that have been made up to that point. There is one more cleanup report: The MARC holdings records that weren’t logically deleted when the bib record is still alive. If these aren’t deleted, they will come across in Unicorn even though there are no holdings for that copy. We might also want to run an unlinked item report. There are now 235,000 unlinked item records. Those without barcodes are being manually deleted. Unlinked reports are refreshed weekly, and they could be sorted by location or call number scheme. If an agency wants to run these, ask Spencer.

Brief records will be created in Unicorn from the remaining unlinked item records, but the entries will be by title, not author, so as to avoid authority problems. These will be shadowed from the public view.

Unclassified periodicals will generate an XX call number, which will look confusing in Unicorn. We could do things in NOTIS now to avoid getting the XX in Unicorn. The copy holdings could be changed to read things such as:

shelve by title v.1

no call number v.1

Journal of [ ] v .1

There is a 40-character limit, and it includes the enum/chron. If you type it once in NOTIS, it populates all of the places that it should be in Unicorn, but if you wait to make the change in Unicorn, you will have to do it manually in each place.

Unicorn has a facility that allows the display of a substitute text in the call number area, but it would have to be a system-wide decision.

Note: Since the meeting, Gary Charbonneau has identified a fix in Unicorn so that all XX call numbers do not display in the WebCat. No cleanup is needed in NOTIS to avoid this problem.

The next meeting will be held Sept. 15 at the Indianapolis Medical Library.

Note: This meeting was subsequently cancelled.

Submitted by Lois M. Sewell