Versions Compared

Old Version 4

changes.mady.by.user Jon Corson-Rikert

Saved on

compared with

New Version 5

changes.mady.by.user Jon Corson-Rikert

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

  • Brown – (Ted) – Still cleaning up data that has been loaded; met with people who maintain the local faculty information system and will be getting current faculty positions and titles from that.
    • Also created new local object properties to display service to the university and other types of service in line with how they typically appear on faculty CVs – looking at how to display date ranges using custom list views
  • Buffalo – Buffalo – (Mark) Has VIVO installed on a new production server and configuing mod-jk
  • Colorado --
  • Cornell --
  • Duke --
  • Florida –
  • Duke – (Richard) Upgraded test environment to VIVO 1.5 but had a question about exporting a large model – kept running into memory problems, so tried exporting data in chunks by getting a handle on a dataset vs. a model – there are a few ways to get a query in there.  Brian – definitely want to use the actual SDB dataset object, not Jena models, or will miss out on the optimization that SDB does for limit queries.
    • Also found that when had been using named graphs, with 1.5 the graphs have to have real URIs, not just strings.  Modified the graph names in the SQL table and reactivated the graph again, and seemed to work. Brian – Yes, going forward the graph names have to be valid URIs
  • Florida – (Matt) Fixed the indexing problem – had to use a Java debugger and go through the SQL tables to find the source; a manual index would fail, and incremental indexes had not been working for a while.  A single form feed character pasted into VIVO from a PDF in March of 2011; cannot reproduce the error by pasting the same file back into more recent versions
    • Now working on getting the logs to reflect manual edits, that previous versions used to log in the vivo.all.log – Brian suggests modifying the auditor, since none of the logging functionality is built into VIVO that was locally modified at UF; with 1.5 have one central place to listen to all the changes via the new RDF API, and the image editing must not yet be using that which is why it still shows up in the logs.  By switching to logging from the auditor in the RDF API, you should see all the edits again, but will have to filter out the inferred triples that are also inserted by the application at the same time. (See dev list messages)
  • Indiana -- 
  • Johns Hopkins --
  • Memorial University of Newfoundland --Newfoundland – (Lisa) 
  • NYU – (Yin) Looking at the VIVO development model and came up with a nice way of saying synchronized with the development master branch as checking in local changes to the search indexing and results display. Substantially replacing some sections of code as converting VIVO to search people only – redefining how the indexing and the weighting goes, using what is now still a prototype; now wants to be able to keep up with whatever changes continue in the dev version so won't have to continually. Will write it up and solicit feedback from others in comparison.
    • could be straightforward to swap out the way that indexing is done for people, but less clear how alternative indexing would work on other types of entities like publications, events, organizations, etc.
    • is there a branching model that is recommended?  Jim – holding pretty close to the Gitflow model; Yin – use that in-house already
  • Scripps --
  • Stony Brook – (Tammy) Erich has WebID working on sample code and still looking to bring into VIVO and will contribute the module.  Upgraded to the latest Tomcat 7 and JDK 7 and Erich is working with both of those.
  • UCSF – (Eric) Getting ready to release the RDF version of Profiles, that is compatible with the VIVO ontology, this weekend.  Will be doing a panel presentation on OpenSocial at the upcoming AMIA conference.  Not sure whether anybody else has gone live with the OpenSocial additions to VIVO?  Alex – has discussed features like the slideshare gadget at some meetings and gotten positive feedback. Eric – reviewers for the panel liked the concept of bringing content in from industry.  Right now are storing the data in a relational model but now sees how to store RDF from working with the RDF version of Profiles – wondering what should be in RDF vs. relational – some content wants to be available as linked data, while other data might be more private and/or just transactional support.  Has there been any thinking about that on the development team?
    • Jim – so far all the discussion has involved in storing as RDF; Jon – VIVO user accounts are stored in RDB, a different data store from where the public data is stored (SDB)
    • Eric – Loki stores everything in RDB and exports on request as RDF for linked open data; Profiles is more of a hybrid, in part due to its transition from a purely relational product, but also perhaps for performance reasons and because it uses a .Net stack. Profiles has kept passwords in RDB, for example.
  • Washington University – (Kristi) Interested in discussions about the Implementation FestWeill --
  • Weill – (Paul) Still working to a deadline of getting publications clean enough to be used as the basis for a separate faculty reporting exercise. Are making very specific Scopus queries for each faculty member rather than relying on general Harvester queries.
  • other --

Other updates or topics

  • Implementation Fest dates and agenda
    • still working on reserving spaces, which may be easier to schedule on Thursday or Friday at the Law School where the I-Fest was held last year
      • could then potentially have recreation like skiing on Saturday
    • may check with a few others spaces including LASP
    • hotel in walking distance from the Law School has a rate of $84/night
    • might be able to do a code sprint with a smaller group at a local startup; open source projects have sometimes done that
    • we'll look for good snow ...
  • Visualization sprint at Indiana
  • Upcoming presentation February 7 on Karma, a data integration tool used to map USC faculty data to VIVO
    • Powerpoint from 2012 VIVO Conference presentation by Pedro Szekely of the USC Information Sciences Institute

...