Meeting Minutes for 06/02/2020

Discovery Committee Meeting
June 2, 2020
Minutes

Demo of new features, bug fixes, and documentation for the upcoming release

• Release Notes

Grouping Project Progress (Minute 12:38)

• Pascal pointed out the Record Grouping Testing Servers information on page 2 of the Grouping Project Progress documentation so that you can browse to examine and evaluate your recordsets.
• During the month of May, work was done on improving movie grouping by replacing the grouping author with runtime.
• The runtime will be round to the nearest 10 minute interval and is located in the Marc record 008 playtime positions 18-20.
• There is a sizable number of records that do not populate the 008 fixed field playtime information.
• There are a few coding issues to point out: 
  • Data not populated :  Empty, |||, ---, nnn
  • Erroneous: other non-numeric codes which can happen when the positioning is off on the 008
  • 000: denotes playtime exceeds 999 mins
• In all those situations an attempt is made to parse the physical description in the form of:   
  • XX hrs., XX min
    • XX here denotes a length of numeric characters
    • Matching ‘hr’ or ‘hrs’
  • XX min
    • XX here denotes a length of numeric characters
• A Cataloging Collective group has been evaluating the playtime movie grouping  
• It would be good if everyone on the Discovery Committee could also evaluate the movie grouping using playtime from the grouping server for their library.
• If you have a movie that is a director’s cut that is more than a 10-minute difference from the original movie, those two will not group automatically.  They will need to be grouped manually. Note: some titles can be closer than 10 minutes apart and still not get grouped together (e.g. if one is 86 minutes and the other is 83 minutes they will get rounded to opposite numbers). The amount of time came from an analysis of a smaller amount of time that did not work as well as the 10 minutes.
• Typically, any movie with multiple productions (i.e. Swan Lake)  that have similar playtimes will need to be ungrouped.  Also, movie remakes may have the same playtime and will need to be ungrouped.
• Improvements with Hoopla movies that matched up with their physical components.
• Bad format determination will result in a bad grouping category determination which will lead to bad grouping. 
• Pascal pointed out stats for the grouping test server:
  • 233631 works with the category of movie that also have playtime for the grouping author
  • 9909 works with the category movie have an empty grouping author
  • 120 works with the category movie have a regular grouping author  (These are Marmot’s Overdrive streaming videos)
• OverDrive does not provide runtime information through the metadata APIs. Marmot has put in a request for an enhancement to include this information, but cannot say when this might be resolved.  
• Any OverDrive streaming videos will not group up with their counterparts from other sources.  You will need to manually group these titles.
• Please use the grouping test server to do searches and noting any issues you find. The Pika team would like to hear your overall view of the movie grouping.
• Title normalization bug 
  • Normalization of grouping title goes awry for the situation of title ending in "by {authorName}"
• Project Wrap Up plan
  • The team is switching to quarterly releases and the next release will be the last for the grouping project.
  • The plan is to use the month of June to finish up any new developments or improvements on the grouping algorithm.
  • In July, the grouping algorithm will be moved to the test servers to find any issues.
  • August will be used for a concrete solid plan for migrating the production server from the currently existing grouping algorithm to the new one that is currently on the testing grouping server.  

 Quarterly Releases going forward (Minute 37:56)

• The developers are moving to a quarterly release cycle.
• The next release will take place in September.
• This new release cycle is to allow developers to have a little more time between releases to be really able to focus on their development work.  
• The monthly release schedule and all the work that goes into doing a release wrapup, as well as pushing the release into production is fairly time-consuming.  The quarterly release will give them more time developing instead of doing a release on a monthly basis.
• The Discovery Committee meetings will also be moving into a quarterly structure.
• When the Discovery Committee next meets in September there will be new activities available for group participation.
• The group will be asked for their feedback and prioritization on an ongoing basis for new feature requests to avoid a backlog.
• C.J. will be spending 50% of his time working on VuFind for the academics and that work will be contributed to the Villanova VuFind code base. He will also be working on a roadmap for our academic members.
•  Marmot is leading the public library SIG for the Folio group to begin a gap analysis with several other libraries across the United States. Marmot will be turning to their members and the LION Discovery Partner for help with this special interest group.  This group will help to identify some of the work that needs to be done for a public library-oriented version of Folio.
• Marmot will be looking at Folio in its current form for the academics libraries, as well as to see what role they might play with an eye towards offering their academics and public libraries a choice of systems and discovery layers.

Discussion of development priorities for the next release (Minute 43:27)

• Road Map
• The new quarterly release means Marmot will not be putting as much time into Pika. It will allow more time throughout the quarter to work on large projects that go along with Pika, as well as improving Pika. 

Current Release

• Implement RBDigital magazine APIs
  • There are multiple issues with the API structure on RBdigital end that has made this a more difficult than originally anticipated.
  • Some items are being rolled back to start the process over
  • This should be ready to go in with the full implementation in September
• Stability Version - Record Grouping Improvements
  • Still a number of specific tasks needed for the improvements
  • Pascal will be taking the rest of the summer to make sure things are working properly before this goes into production.
• Admin user options project
  • C.J. will continue to work on these options
• Investigate Marmot Folio Implementation
  • Chris will be working on this project broadly

Other discussion topics (Minute 45:58)

• Pika configuration changes for reopening processes
  • A lot of changes have been made to your Pika configurations in the early wake of the pandemic
  • Ashley has been in touch with some libraries to help them make changes as everyone is moving to new processes and reopening procedures.
  • Ashley offered to help any library with their Pika reopening changes.  She also reminded people that Pika may not be able to do everything you need during this time.   
  • If you need assistance please contact Ashley at pika@marmot.org

Next Meeting is Tuesday, September 1, 2020