The LD4L-Labs and LD4P grants looked at various aspects of querying, searching, retrieving, and integrating external information from sources such as the Library of Congress, Geonames, and others. In these projects, we reviewed how to use and integrate the Questioning Authority framework and infrastructure developed by Lynette Rayle and Dave Eichmann.  Below are links to and information about the work we completed in the last grant regarding lookups (and this page may also serve as a way of bringing together information about the work we will do in the LD4P2 grant but that remains to be decided).

LD4L-Labs + LD4P (but also relevant to LD4P2)

• Questioning Authority
  
  • Architecture for Authority Lookup : This page provides information about architecture, configuration, and has multiple important links. I'm including a few here as well.
    
    • Mockups for Lookup with Context
    • http://lookup.ld4l.org/authority_list : Showing all the authority QA endpoints as well as providing information about status
      • For reference, the list of authorities currently implemented in QA includes: Agrovoc, DBPedia, Geonames, Getty, LOC including names, genres, subjects, MESH, NALT, OCLCFast.  Different versions may exist between cached versions and non-cached or direct lookups (Lynette, feel free to correct this)
      • Current status of contextual information is that information at the first level (i.e. subject predicate object in the returned triples) is returned as part of the context object.  Additional work will need to be done to refactor the context code and expand the context that is returned. 
• Lookup sources and information: Which sources do we lookup? (Also see LD4P2 work below)
  
  • For LD4P2, we are still discussing whether/how existing works and content from sources like OCLC/LOC and Shared-VDE should all be made available through the same QA framework as used for authorities and external vocabularies.  In LD4L-Labs and LD4P, we discussed which lookups we needed to integrate and for what there should be support.
    
    • Some discussion has started regarding which lookups will be relevant to Sinopia
• Cataloger interactions with/feedback on lookups
  
  • As part of their current MARC-based workflows
    
    • Tracey Snyder and Ardeen did excellent work capturing and describing information about how they conduct searches in their current systems

    • "Here is a two-part demo of how I search for authority records for LCGFT genre/form terms using OCLC Connexion. Note that there are also more sophisticated searches that one can do by combining keyword searches that search multiple indexes at the same time — I do this all the time when I am combining keywords from a person's name with keywords from the title of a work, searching for the name-title authority record for a specific work by a specific person. This demo only searches a single index, the Genre index, because that is our immediate purpose. Note that Connexion allows AND, OR, or NOT operators. If I enter all my keywords in a single field searching a single index, an AND search is performed."

      • Tracey and LCFGT
        

        • https://www.screencast.com/t/V4NPz8HnHc

        • https://www.screencast.com/t/2CsbnqycN

      • Ardeen and Voyager
        
        • https://www.screencast.com/t/54ddbs5Nd
  • Feedback from evaluations/usability studies
    • Round 2 and 3 of usability studies on VitroLib includes reaction to incorporation of lookups
      
      • https://wiki.duraspace.org/download/attachments/77447843/vitrolib-usability-round2-summary.pdf?version=1&modificationDate=1491579091313&api=v2 (Round 2, specifically item 6)
      • https://docs.google.com/document/d/1IDGt1ZRMuSl15NehNFIm4EMK8kBdcxNO_ioG7cQZEbw/edit?usp=sharing (Round 3, several observations center around easier ways to pick often-used items from dropdowns, being able to view the authority record directly, being able to pick between person/organization type in LCNAF search, etc.)
      • I will also look back at lookup-specific or context-specific feedback in the individual results again (in a brief re-review) to see if more can be gleaned
• Lookup: Testing and evaluating search results
  • QA - LCGFT Test Queries
  • Configuration for testing: https://github.com/cul-it/qa_server/blob/master/config/authorities/linked_data/scenarios/locgenres_ld4l_cache_validation.yml

LD4P2

• Brainstorming Doc: https://docs.google.com/document/d/1uKcYhWwVDcvoCm1x_7CXsw1ooJH4F3z6dRhUtG0wi2Q/edit?usp=sharing
  • This represents a discussion around what work may be done or relevant but encompasses many more items and questions than we know we will be able to address in LD4P2. 
• Which lookups?
  • Based on work started in the previous grant, the following spreadsheet is being used to track which lookup sources should be considered for the Sinopia MVP
    • https://docs.google.com/spreadsheets/d/1rPvEoP9iYNkxJ0eWC8gXe3ci7e6mhW0da59xkGhadi0/edit?usp=sharing
• Lookup performance and load
  
  • Simeon's review of cataloger lookup actions within sessions and analysis of expected lookup load : https://github.com/LD4P/ld4p2-cornell/wiki/Expected-lookup-load-from-cataloging