D&D Minutes 10/29/09

D&D October 29, 2009
Present: Peter, Dan, Shawn, Robin, Scott, Lyman, Elizabeth, Donna

1. Criteria Next OPAC
Should next OPAC be able to pull in CDI and Special Collections records
along with Voyager records—yes, and it should look good doing it. Higher
priority.
Should the next OPAC be able to search external databases too?—yes, and
look good doing it. Lower priority. How about CRL?

2. Hathi Trust (260,000 records)
Catalogers' response raised concern about difficulties with leading
articles in titles. Lyman said these could be fixed. Some issues with
data being located in different fields than what we use, but we could
move these if it's a problem for patrons.
Amounts to a 17% increase in size of our database, with old material. If
it turns out that these records drown out records to our original
collection, we'll have to delete them. Loading will take time. Unloading
will take time if we have to do it.
Catalogers are concerned about degrading the indexes in Voyager, which
may refer to authority control. It may be that the Hathi Trust records'
date format will cause them to not sort properly.
Next step is for Lyman to load 1000 records into Voyager and see how
that works.
We'll try it and at our next meeting we'll work on a survey.

3. LinkSource report
Linksource fails to get to the article at all 20% of the time. 19% of
the time you get to the A to Z list and can work your way to the article
with effort. Our theory is that linksource's database has many variant
titles for journals.
Another problem is that we can't successfully prioritize the vendors.
Linksource seems to override this ranking and pushes us to EJS.
Identified problems:
1. Linksource fails (20%), perhaps because database is not clean.
2. Linksource sends you to the A to Z list (19%). This should never happen.
3. Linksource doesn't consistently accept our prioritizing of vendors,
and often sends users to EJS instead of the publisher.
4. Linksource customer service has been unable to fix problems 2 and 3,
and is sometimes slow to correct errors in their databases.
5. Linksource's intermediate page is difficult to customize and make
attractive.
6. We'd like more functionality, for example, we'd like to turn on parts
of Lexis/Nexis and turn of others.

We can send the report on to the Dean's council with a cover letter from
Peter.

4. Commenting in the OPAC
Not used much at other locations

Minutes 10/15/2009

Present: Laura Haines, Dan DeSanto, Robin Katz, Lyman Ross, Scott Schaffer, Shawn Biegen, Fran Delwiche, Selene Colburn, Peter Spitzform

• The Linksource subcommittee has finished; a report is forthcoming. Tina mentioned that an EBSCO rep had informed her that a new version of Linksource had been released, and is in use at UVM, with many alleged improvements.
  

• 260,000 HATHI truest records are now available and can be uploaded to Voyager. The records contain minimal cataloging. OCLC is working to create fuller cataloging for these records, but it's unclear what the status of the project is, or what the cost will be. Lyman will upload a sample batch of 100-200 records, and Peter will share these with cataloging librarians. D & D supports the project; we're interested in getting something added to the record, such as [electronic resource] or [electronic book], so that when it comes up as a duplicate in search results, the user will understand the difference in format.
• We looked briefly at some Vu Find implementations, then talked about the process for evaluating and recommending "next generation" solutions. Two subcommittees were formed: one will create a comprehensive list of possible solutions, reporting back in two weeks (Scott, Robin, Peter); a second will create a rubric/list of evaluative criteria (Lyman, Selene, and up to two others -- to be recruited by Peter), reporting back in four weeks.

10/1/09 D&D Minutes

D&D October 1, 2009
Present: Peter, Dan, Shawn, Robin, Laura, Fran, Scott, Daisy, Lyman,
Selene, Donna

1. Linksource sampling
Using the find by citation search feature: 75 citations tested by each
of the 4 testers
If success is defined as: Got to either the article, the issue, or at
least the journal with full citation: Scott found 77% worked. In his
opinion the majority of the problems were non-scholarly things (eg
newspapers) in Lexis-Nexis. Also some problem with Open Access. 40-50%
went to the article level.

Shawn was checking title search, title plus ISSN, full citation search.
Problems found: linksource linking back to the Title on the a to z page.
Lyman and Laura saw that too.

Lyman found that Lexis Nexis and Gale did not work very well.

Laura had 41% that were "successful" as defined above.

Lexis nexis may have improved more recently.

They will be meeting soon to aggregate their findings.

What are we going to compare these percentages against? Is this the
industry standard or could we do better with another vendor? Is there
anything in the literature?

Linksource does provide for a way to skip less reliable vendors, but
Shawn reports that this feature often doesn't work.

2. Relevancy Ranked Search
Why didn't we implement this? Let's take a look at this and see if we
can just provide it as an option in the production catalog?

3. Downloading Hathi Trust records
A lot of work. Is there enough interest in this to justify Lyman's time?

4. More info in stacks on what subjects the call numbers correspond to.
We like the idea. The signage project is currently in process. Maybe
this is a good time to pilot something? Maybe Trina can talk to Access
Services?

5. Next OPAC
Started reviewing desirable traits for our next OPAC
Icons to distinguish one type of material from another—we have a few,
we'll keep an eye out when we look at other systems to see if this is
handled better elsewhere for books, journals, and gov docs and
electronic resources.
Boston College Aleph catalog has a locate button with a map.
http://library.bc.edu
From the ASU library catalog you can click on a button and rerun the
search in the CRL catalog.
We'll continue this discussion next week.

--