Time/Place

This meeting is a hybrid teleconference and IRC chat. Anyone is welcome to join...here's the info:

• Time: 11:00am Eastern Daylight Time US (UTC-4)
• ReadyTalk:
  
  • U.S.A/Canada toll free: 866-740-1260, participant code: 2257295
  • International toll free:
    
    • http://www.readytalk.com/intl
    • Use the above link and input 2257295 and the country you are calling from to get your  country's toll-free dial-in number
    • Once on the call, enter participant code 2257295
• IRC:
  
  • Join the #duraspace-ff chat room via Freenode Web IRC (enter a unique nick)
  • Or point your IRC client to #duraspace-ff on irc.freenode.net

Attendees

• Greg Jansen
• Esme Cowles
• Ye Cao
• Markus Haarländer
• Matthias Walter
• Eric James
• Michael Durbin 
• A. Soroka

Agenda

1. 20 - Review current fcrepo4 work.
   
   1. Triple store integration
   2. Customizable search indexes
   3. Large file performance
   4. Fcrepo3 migration
   5. roles-based AuthZ
2. 10 - Scalability of the API for children
   
   1. Moving away from using Jena Model as container for triples in the kernel
   2. Moving to a streaming API all the way through the kernel
3. 10 - Stakeholder needs in a customizable search indexer (Solr)
4. 15 - Wiring and configuration (see also Comparison - Wiring)

Previous Actions

• Michael Durbin to create Pivotal tickets for UI polish
• Benjamin Armintor to send email for others to sign his and other committers' keys

Minutes

1. reports on current work
   
   1. triple store integration  - Esme 
      
      1. will be looking at Mulgara and JMS indexer to incorporate non standard SPARQL queries
      2. looking at multiple levels based on reviewing use cases
      3. kitchen sink now has triple store  ( Question for Andrew:  Is vm built from master automagically )
         1. Answer from Andrew: "Yes, Jenkins deploys kitchen-sink on each commit to: http://futures6.fcrepo.org:8080/fcrepo/rest/"
   2. large file performance - Eric  
      
      1. 3 avenues explored
      2. researching different configuration settings ( e.g. caching hashes showed dramatic improvement )
      3. Scott commented on how fixity done in Fedora3
      4. fyi, looks like we actually read the data and do a new checksum when we do the fcr:fixity check, the low-level code that does the reading/checksumming is here:https://github.com/futures/fcrepo4/blob/master/fcrepo-kernel/src/main/java/org/fcrepo/kernel/utils/LowLevelCacheEntry.java#L92 (Esme)
   3. migration - Mike
      
      1. 90% completed and the remaining 10% is slow
      2. working on display, especially focused on performance when many children as displayed
   4. roles based - Greg
      
      1. will be working on integration test needed for effective roles for tree object
      2. roles can have expanded permission set ( Scott )
      3. XACML will be used for each role
      4. policy enforcement point
2. API for children
   
   1. effective hierarchy needed to maintain good performance
   2. a couple of tests suggested  (Greg has the action items)
3. search indexer - Markus
   
   1. external indexer (esp. Solr)
   2. Design - Customizable Search Index
   3. Scott described how they use Solr in F3.   change METS to Solr XSLT to customize
   4. what should be indexed?   data streams, text extraction from spreadsheet, PDF, etc.
   5. Possibly use METS which can point to binary files
   6. maybe use a default Solr schema; consider exposing to public and document how to use
   7. alternatives to XSLT should be considered
   8. need to keep the idea of distributed hardware ( e.g. separate Solr and Fedora)
4. wiring - Adam
   
   1. changes will be hard and are hard now
   2. future proofing may not be as important as buy in from the group
   3. examples used currently or in the past: struts, Spring, JAX, CDI
   4. feedback from the community needed for https://wiki.duraspace.org/display/FF/Comparison+-+Wiring

Actions