Huda Khan - Schedule usability work. Will forward tasks draft to group and to Kevin/Usability Working group today
2020-06-12 Met with Kevin and usability group on Monday. Agreement to do think-aloud and have reps interested in participating, might also get folks from access services. Plan to get tests done by June 25 giving time to create a report
2020-06-22 Three usability tests scheduled Monday + Tuesday
2020-06-26 Five usability tests completed, usability report forthcoming
Steven: Report outs in metadata group included discussion of usability tests. People who use online search a lot (search engines etc.) seemed to understand auto-suggest functionality quite well.
Huda Khan - To make slides and such to go through indexing approach, compared with client-side only, lessons learned, etc.
What about cases with many variants? All are added to the bucket field but only those matching the query are shown as `aka`
Why have a combined bucket field rather than searching multiple fields? Provides simple way to search although current implementation searches label and bucket fields to enable matches on the label to be boosted. Also the bucket field has special matching criteria such as start of word, which isn't the case for all fields we might be interested in
2020-06-26 Slides to be completed early next week as part of documentation for project
Huda: Wrap up work. Working on some remaining bugs, hoping to address by Tuesday (and if not, will document what remains to be done). Lessons learned draft started. Also brief indexing script documentation (for how to run the script). WHAM! demo video segments to be done by: indexing (Huda - Friday), controller front-end (Tim - Monday), packaging (John - Monday).
John worked on packaging, is moving Tim's code into nectarguide. Have decide not to adopt to use twitter typeahead but instead a more direct move.
Last week: Lynette has mostly been focusing on exhibits. Dave is working on performance improvements to eliminate SPARQL quest after matches for LOC. Still have Steven's additional tests, want to see how capitalization tests work and should have same results – hope to get these set up this week
This week's update:
Accuracy testing: 2020-06-24 Accuracy Testing Focused on Capitalization Observations: Stemming is not happening. Exact matches aren't found when the preferred label is language tagged. Dave is looking into turning on stemming and exploring how to handle labels with language tags.
Performance: Dave is reindexing LOC to use the new indexing that keeps the extended context in the index.
First use cases added around looking up authorities: keyword or left-anchored search. Filtering also mentioned as a desired feature. Other use cases included caching, batch processing both automatically and supervised (e.g. reconciliation).
Issue raised that perhaps we should first define a model that would be supported by shared APIs even if this data stored locally in a different data model.
E.g. If requesting information about a person entity, data is returned using the same Person ontology regardless of data source/authority.
Discussions will continue around use cases and how the working group will address them.
Possible relationship to entity management work in FOLIO
How much full LD vs linky-MARC?
CUL LTS working group is hoping to add URIs to MARC before FOLIO migration as a substitute for some of the heading management we do in Voyager in other ways. Perhaps won't be done ahead of switch but as part of the data migration. Considering how to update/maintain going forward
2020-06-05 Agree that doing some analysis/write-up for the Cornell context would be good, ACTION for July
Greg has been working on setting up a copy of Cornell Blacklight
Thursday meeting - went well, imagine having follow-ons to flesh out the work packages, consider engagement and how the target for cohort/PCC has shifted – with more PCC involvement we may narrow focus more toward monographs and serials. We should be careful to have shared understanding of scope. Discussion toward the end of to-dos for the specific work packages, likely need a meeting for each. Need to understand collaboration between Stanford and Cornell on discovery. Not sure of link (if any) between Sinopia data and discovery work
Steven doing workshop in week of July 9: details here: https://github.com/sfolsom/LD4_2020_Conference_Linked_Data_WorkShop. Comments welcome, but no pressure. I’m still working with the Conference Organizers to firm up workshop slack, email, and office hour communication.
Everyone encourages to respond promptly to conference organizers
Discovery track: July 20-23. Huda + Astrid tentatively scheduled for July 20 and going live.
Lynette and Jeremy - July 9 doing a recording with live Q&A
Lynette no longer doing presentation of best practices for authorities WG - too early in process