DC_Fedora_User_Group_2014-03-10.pdf

What

D.C. Fedora User Group Meeting

Where

National Museum of the American Indian
Washington, D.C.

Room 4018, which is a space just off the main hall, so no need for special access

When

10 Mar 2014
10:15 am until 5:00 pm

Attendees

Agenda/Presentations

Summary

10:15 – Welcome and introductions all around

• Thorny Staples - Smithsonian
• Andrew Woods - Fedora Tech Lead
• David Wilcox - Fedora Product Manager
• Jonathan Markow - DuraSpace
• Tom Cramer, Stanford
• Ron Jantz, Rutgers
• Don Gorley, National Agricultural Library USDA
  
  • Ursula Pieper
  • Wei Wu
  • Chuck Schoppet
• Rob Cartolano, Columbia
• Bria Parker - NASA Goddard Library, Space Flight Library
  • Kevin Rice
  • Mitzy Cole
• Stefano - Art Institute of Chicago - new Fedora adopter
• Patrick - Northeastern University
• Fran Stern - Smithsonian
• Ti Amy - NLM, long-time user of Fedora
• Jenny - UMD
  • Ben Wallburg
• Adam Soroka - UVa
  • Mike Durbin - UVa - repository manager
• Josh Wesgard - UMD

10:30 – 10:50 David Wilcox – update on the Fedora community and DuraSpace

• slide show
• revitalized technology
• revitalized pool of contributors
• increase community involvement in project
• use-case driven
• review of steering and advisory group
• Fedora Community
  • 326 Registered Fedora Implementations
  • 21 new instances in 2013
  • 858 members of fedora community mailing list
  • 40 Fedora Sponsors
  • 19 Active Developers
  • 17 Members of Fedora Advisory Group
  • 10 Members of Fedora Steering Group

10:50 – 11:20 Andrew Woods – update on Fedora 4 development

• Fedora 4
• call to engage in generating use cases, testing
• Fedora 4 Features
  • Content Modeling
    • nested or hierarchical structure
    • validation - define properties on objects
  • Authorization
    • application has pluggable authorization mechanism
  • Durable storage
  • Versioning
  • Scale (large files and many files)
  • Linked data / RDF (and external triplestore)
    • although not part of core base
    • important that triplestore be readily available for Fedora
    • pattern for triple store and SOLR index
      • every event that takes place on repository, JMS events come out of repository
      • completely functional message consumer
  • Internal and external search
  • Transactions
    • any action, largest bottleneck is persistence of action
    • when save takes place, takes significant part of action's time
    • pull together a series of actions into a single transaction
      • one larger save at the end
  • Performance
    • recent sprint
    • 30% faster using transactions for update operations
  • Clustering
    • consistent, scriptable way to put together Fedora servers

F4 Timeline

• Spring 2014 - 3.7.2 release
  • around code4lib conference
• Spring 2014 - 4.0-beta release
  • feature complete beta in spring Q1/Q2 boundary
• Engagement with the community
• Beta acceptance testing
  • download the alpha and beta, to provide feedback
• Beta pilot projects
• 4.0 Fall 2014
• Mailing Lists
  • fedora-community
  • fedora-tech
  • dc-fedora-users

Questions:

1. Migration path from Fedora 3 to Fedora 4
   • greenfield first
   • options
     • projection of Fedora 3 content in Fedora 4
       • can use immediately
       • copy over time
2. Ron Jantz
   • is there a 3.8?
   • No, we do not expect a 3.8
3. Jonathan
   • acceptance testing using F3 connector?
     • not hardened code yet
   • F4 has one click install, amazingly simple
   • what can we do to lower the bar to get you to test it
4. Transactions
   • we rely on transaction msg that they've actually occurred
   • at end of commit, want JMS msg, that transaction is complete
   • Adam - answers the question

11:20 – 11:50 Bria Parker and Kevin Rice, NASA Goddard Space Flight Center Library – update

• Bria Parker - metadata librarian - since 2010
• Kevin Rice - web programmer since 2012
• Mitzy Cole
  • contractors at Goddard Library
  • they've been downsized
  • originally 2 programming, 1 metadata librarian
• Godard Library Repository
  • http://gsfcir.gsfc.nasa.gov
  • Drupal 6 on top of Fedora 3.3
  • harvest, scripting before it gets to Fedora
  • two collections
    • JSP scripts load collections from outside Fedora
    • two collections entered right into Fedora using Drupal interface
• what's in it?
  • colloquia, authors and publications
    • haven't ingested PDF
    • use link resolver to get to PDF content
  • separate authority author objects
  • balloon technology
    • scientific documents
  • a list of publications,
    • RDF - all other Goddard authors
    • all happening in Fedora
  • script to check author authority match
  • plans to move to Drupal 7
    • original people that built it are gone
    • some documentation
    • some logic is lost
• Kevin Rice
  • Drupal 6 - piecemeal version of Islandora
  • works in a similar process to Islandora
  • do not intake data through Drupal
  • only display it through Drupal
  • migrating from drupal 6 to 7
    • mysql transactions have changed
    • take a step back
  • fixing Drupal issues
    • will help get to new versions of Fedora
  • change made to core Fedora system
    • by local developer
    • no feature needs for future versions of Fedora
  • harvest publication information from variety of commercial databases
    • since we started, VIVO came out! 8-)

11:50 – 1:00 Lunch on your own

1:20PM - DuraSpace - Jonathan Markow

• overview of chart
• Sponsorship vs. Membership model
  • "Sponsorships" not part of European approach/budgets
  • many are budgeted for "memberships"
  • libraries have "membership" budget lines
• membership - more participatory governance

1. Q:Rob Jantz
   • challenge with synchronizing our development with Fedora development

1:50 – 2:20 Ron Jantz, Rutgers University – update on RUcore and their research data portal

• DOI and micro-citation
• RUCore Data Repository
  • Research Data Working Group
    • Established group including library liaisons
    • group provides supports for grants and NSF Data mgmt plan
    • Developed extensive metadata profile for research data
  • Research Data Portal - recent additions
    • ingest of multi-level directories as submitted by the researcher
• link data to articles and vice-versa
• Rutgers allows a researcher to give us a full multi-level directory all metadata is typically in the file name
• using CDL's EZID service
• 35,000 objects in RUCore
• Questions
  • Do we want to assign every one of those items a DOI?
  • we think probably not, but then how do we filter out?
  • Not everything in RUCore is research data
• Example: 3,000 Roman coins
  • we take 7 images per coin
  • we would probably not create a DOI for every image
• can we do even finer grained linking, subset of a dataset, paragraph in a book, video clip, etc.?

1. Questions:
   1. Robin
      • ORCID - give them a choice
      • Ron
        • fair amount of faculty pushback
        • Universities trying to do University-wide
          • don't want to fill out "yet another profile"
        • some faculty pushback
   2. Tom
      • upload a whole directory of files, is it one fedora object?
        • Yes
   3. Any migration to other formats?
      • Ron - we do. We take Word documents, migrate to XML, keep original

2:10 – 2:30 Wayne Strickland, Information International Associates – Update on the Federal Science Repository Service.

• Federal Science Repository Service
  • Wayne Strickland
  • Gail Hodge - private contract - public/private partnership
  • Don Hagen - Associate Director (Wayne's boss)
• National Technical Reports Library
  • NOAA - deep water horizon project
• FSRS
  • http://www.ntis.gov/fsrs/
  • by statute, used to charge for service
  • this is changing
  • use our approach with FSRS
• Technical reports used to be much more important
  • Data needs to accessible
• Object model
  • geospatial data example
• NOAA Repository - Deepwater Horizon Repository
  • 8,000 metadata records
• Two Islandora implementations for programs in D of Commerce, and DoD
• Don Hagen - talking with OSTP
  • open source, open access
  • part of the culture shift

3:00PM - 3:30 Thorny Staples, Smithsonian – update on SIdora, a research support environment

• Putting data that has never been in an organized place is hopeless
  • Rather, create a workspace to better manage research data from the start
  • So that when the research process is done
• Research Project - no formal hierarchy, it's a graph
• Two RDF relationships
• We will collect more and more standards
  • concepts are metadata objects
• Discovery and Collecting Environment - starts collecting data right at that moment. Starts gathering links to things, with links to items already in their project.
• DataSet concept - concept of sets to bound the data
• Analysis Environment:
  • Galaxy - workflow management system, by genomics
    • reflect a SIdora set as a Galaxy set, and all the tools in Galaxy can be used
  • Taverna
    • workflow environment, comes with all the R tools
    • can convert a SIdora set to a Taverna set
• Researchers have their own stuff, own tools on their own desktop/laptop
• Have files look like local filesystem (like dropbox)
• Questions
  1. Tom - access
     • Thorny - wants to use disseminators
       • for tabular data files - the idea is when you upload tabular data, assign to a code book
       • don't upload unless you assign the variables
     • can build disseminators on the dataset concept
  2. Do you expect to assign workflows as objects?
     • Workflow in Taverna is already an XML file
     • Adam Soroka - get connected to Kepler
  3. rtc - enforce requirement to upload codebook?
     • Thorny - yes
     • we will ask for absolute minimum metadata
       • try to fill out the code book and ask them to correct it
       • huge payoff when you upload 1,000 data sets with same codebook
       • can generate spss output for use by R later in workflo
  4. Working on relationship with Oakridge
     • Smithsonian, public-facing front end
     • grid ftp in a lab
     • lab's work becomes a research project
     • each

3:40 - Don Gorley, National Agricultural Library

• OSTP mandated that agencies that meet a threshold need to make results of research publicly accessible
• Saw our area as the place to make materials publicly available
• 2 Fedora instances, and ILS
• be prepared to scale up discovery and access
• Single Fedora instance, with 4 to 5 million objects
  • MODS data streams
• A lot of overlap
• A lot like a PubMed Central system
• Use Islandora as the management platform
• For discovery, it's really working against SOLR index, and not Fedora at all
• staff were reluctant to mix
• separate out discovery interface completely from Fedora
  • Fedora is back-end
• build indexes and content servers
  • java applicaion servers
  • nginx front-end
  • SOLR index
  • filestore of content that we are delivering
  • simple pear tree directory structure
• JBOSS domain controller
• SOLR master server - keep SOLR slaves up-to-date
• Content server - copying
• changes in Fedora repository will change
  • Fedora master
  • content server
  • use Rsync to distribute to front-end

3:45PM Ben Wallberg, UMD

• moved toward management
• building up the development staff
• Fedora - many home-grown applications
  • search service - moving to SOLR
• use plain file system storage
  • research hadoop as a back-end for Fedora
  • hadoop has big growing community
  • if we could use hadoop to store large files
  • share results on code4lib
• four home developed interfaces to Fedora
  • 1 staff facing
  • 3 public facing
• open source - hippo
  • libraries web site
  • use it for other interfaces
• focus on improving loading and backlog of collections
• we are on Fedora 2.2 instance

4:00PM - National Library of Medicine

• John Doyle
• Fedora, Blacklight
• http://collections.nlm.nih.gov
• 2 million items
• working with serial content
• Running Fedora 3.6
• hoping to go to Fedora 4
• indexcat - 3.7 million items
  • special Fedora instance to handle ingest of older XML data
  • mainly journal articles cited over decades
  • not full-text content
  • purely for preservation

Thorny - Closing Discussion

• Technical Training for Fedora 4
• Fedora 4 Migration meeting

Other Attendees

• Tom Cramer, Stanford University
• Robin Ruggaber, University of Virginia
• Jennie Levine Knies, University of Maryland Libraries