Meeting Minutes for 12/18/2019

• Duplicates team report
  • Discussed DVD grouping again.  Running time seems to be the best second criteria for matching after title.  
  • Pascal says it would be fairly easy to program in a grace period so the running time could be a close match, but not exact.
  • We decided to check a sample of records for how many minutes of grace would be optimal, or if exact time would actually be better. <<ACTION ITEM
  • Also, we will look at if the date in the 046|k is helpful.
  • Jamie wonders if that date is more commonly included in newer records.
• FOLIO update
  • Lloyd has been exploring the idea that it might be possible for a consortium on Folio to make use of the Libris software created in Sweden to allow their libraries to work together.  Chalmers University is currently using Libris with Folio. Chalmers catalogs in Libris and downloads records into their Folio system for circulation. It seems like this is the same functionality we would need for some sort of central cataloging utility that could send records to each library’s instance of Folio.  At least that is one way that a consortium could function on Folio.
  • This morning Lloyd learned that Libris is open source.  There may be a consortium in France that is also using Libris.  He saw a demo of it, and realized that the interface is in Swedish.  There may be a demo of Libris cataloging at WOLFcon in January. Lloyd will attend that if it happens.  If it is being used in France, then it must be possible to create different language options and there may already be an English version.
  • Another issue is that Libris is current only working with linked data, not MARC.  Chalmers is doing all cataloging in linked data. Libris can import and export in MARC, but it does not use MARC internally, and Chalmers is only using linked data in Folio.
  • Lloyd thinks that there is probably an older version of Libris that still supports MARC.  Another option would be to create records in OCLC and SkyRiver then import them to Libris.  Or we could just start cataloging in Linked Data which we probably will at some point anyway.
  • Jamie asks if this would make it possible to switch to Folio sooner.
    • Yes, it might make it possible for a consortia to use Folio even though Folio hasn’t built consortial functions yet.  It is still probably at least a couple of years away regardless.
  • Lloyd will be attending the conference of the World Open Library Foundation in Texas next month.  He will be able to learn more about Folio there.

Completed action items

Discussion Topics

• Subject vs. Topic search in Pika
  • We realized that Pika is creating subject facets is a very strange way.  It is pulling the subject headings in the records apart in a way that doesn’t really make sense.  The way they function is described in this document: https://docs.google.com/document/d/1Ii5JqJXEyEDS0V3BC6XjlSVESrv5u2XZ6XVtTsyMKEg/edit
  • This facet logic attempts to pull the subject headings apart in a way they were not meant to be used and it leads to some incorrect results.
  • For example, the Region Facet takes subfield $z from many 6xx fields and subfield $a from 651, but it ignores the fact that a geographic name is often included in the 610 if it is being treated as a political jurisdiction.  It would not be repeated in the 651, so many geographic headings would be missed by this logic. Including the 610 would mean many corporate names would get in the facet, so that is also not acceptable. It may not be possible to create a functional region facet.
  • The LC Subject Facet only includes 650 headings.  It leaves out personal subjects, corporate subjects, meetings, titles, etc.
  • The Topic / Subject Facet is very problematic.  In the default Pika set up this is labeled as the Subject Facet.  However, it only includes subheadings identified as topical. That means subfield $a from the 600 and 650, and subfield $x from all the 6xx fields.  Similar to the LC Subject Facet it is ignoring many subject headings.
  • The Subjects Facet seems more useful.  It includes all the 6xx fields and subfields.  However, it is not used in the default Pika set up, so most libraries are not actually using it.  Also, it does include the 655 fields which are genres not subjects.
  • Jamie points out that the Subject Facet includes non-LCSH headings.  He would find it more useful if it were limited to LCSH.
  • Pascal says the subject facet works differently.  It does not break up the subfields into separate facets.  It creates facets based on the entire pre-coordinated heading.
  • Jamie mentions that the classic catalog will rotate the subheadings so that every subject heading gets filed under each of its subheadings.  This is a useful function.
  • Jamie points out that the subject facet also includes foreign language subject headings, which he thinks is clutter for his users.  He would prefer if the facet was limited to LCSH headings.
  • There is a separate BISAC facet available for libraries where it would be useful.
  • Lloyd suggests that the Topic/Subject facet is not functional, and should not be the default.  It is probably not at all useful. It would be improved to make the Subject Facet the default, although it could use some improvement.
  • The Subject Facet also includes FAST headings, which are likely to be redundant.
  • The LC Subject Facet is too limited because it only includes 650s.
  • Nina suggests we should take this up with Discovery Committee. << ACTION ITEM
  • Pascal would like us to develop a specific proposal for what these facets should do.  He wants an explanation of the difference between subject and topic.
  • Lloyd thinks that LCSH just doesn’t break down this way.
  • Jamie created a list of what he thought the LC subject search should include, he would send that to Lloyd as a starting point.
  • Jamie asks Pascal what is the “by subject” search searching?
    • It searches all of the facets with subject and topic being most important.
  • Pascal also would like to include someone who uses subject searches regularly on this project.
• Chat asks what is the best way to search for kits in Pika.
  • There is a format facet for kits, so limiting by that would probably be best.
  • We look at how Pika determines that something is a kit, it looks at the Record Type in the leader and Material Cataloged in the 007 field.  Those codes will give an item a format of kit in Pika.
• Lloyd asks Pascal if anything can be done about the way Pika is displaying MARC records in staff view.  They display in apparently random tag order. It makes them very difficult to read.
  • Pascal says it would be difficult to fix this because the new API pulls out the MARC record in several pieces with several separate calls.  
  • He will put in a Pika development ticket, for resorting this display in numerical tag order.
  • The MARC display in Pika only includes fields that are used in Pika, so it does not include the 001 for example.  There are other fields that are not included.
• Regrouping previously ungrouped records in Pika
  • Tammy has been testing grouping and ungrouping on the Pika test server.
  • We turned off all manual grouping and ungrouping for serval titles and allowed them to regroup themselves automatically.  
  • Looking at the Lion King movie, she ungrouped the old movie from the new one, then manually created a new group of for the old one.
  • The result is that the old movie group is the first hit and the new movie group is the second hit when you search “lion king” in Pika.
  • We assume that when the number of check outs on the new group exceeds the old one it will move up, but it is still much more prominent than the previous arrangement. 
  • The automatic group here is the new movie.  Any new purchases of the new movie will automatically group there, however if anyone does buy a new copy of the old cartoon, it would group with the new movie and have to be manually ungrouped and regrouped with the cartoon.
  • Jamie points out that we are thinking about a new method for automatic grouping for movies based on the 046|k which would sort this out.  Implementing that would solve this problem. We could always add an 046|k to force grouping rather than making changes in Pika. 046|k might have to be a protected field.
  • Jamie asks Pascal if there can be a hierarchy of criteria.  It might look at the 046|k first, then if that is not there, then look at running time.
  • Pascal says that would not be easy.
  • 046|k would be easy to add to records in order to control grouping.  There are not that many records where this is a problem.
  • There are rare cases of two movies with the same name made in the same year, so we will still need a manual function.
  • Lloyd is working on analyzing how many minutes grace period in running time is optimal.
  • We want to include 046|k regardless.
  • Jo points out in chat that you can sort Pika results by newest.  When you do that the results get very strange because you lose all relevancy ranking.
  • Tammy asks what happens to manual grouping when we change the grouping algorithm.
  • Pascal says it would change the grouping ID, so we would have to migrate the manual grouping to the new IDs.
  • The timeframe on the grouping algorithm overhaul is several months out.  He will start with language grouping anyway.
  • Chat from Nina: I thought that once things were ungrouped they could not be regrouped again.
  • That’s what Lloyd thought as well, but by experimenting we find that we can.
  • Pascal says there are some situations where they can’t be regrouped, but he is including fixing that in the grouping overhaul.
• Illegal Aliens heading update
  • Not much has happened on that.  Lloyd suggests we want until after the report from ALA Subject Analysis Committee at Midwinter.
• Cataloging standards for brief records ​
  • We already approved this change earlier in the meeting.​
• Jamie points out there is a bug in Sierra where sometimes it is not possible to flip review files to different record types.  It fails when you try to flip to a different record type and also search for a characteristic of the new record type. We think this is fixed in Sierra 5.0.
• We will probably upgrade to 5.0 in January.

New Action Items

Ongoing Action Items

Next Duplicates Sub-committee meeting: Jan 8

Next UCC meeting: Feb 26 (no meeting in January because Lloyd will be at WOLFcon)