Personal tools
  • We're Hiring!

You are here: Home Community Minutes Conference Calls 2010 2010-05-18 Tuesday Meeting

2010-05-18 Tuesday Meeting

Agenda

Remember: Agenda must be complete (with estimated times) on the day before the meeting. Any additions after that must go at the bottom (AOCB)

  1. Accepting minutes from last meeting

  2. Matters Arising (<10 mins)

    • Tagging Group: Meeting on Friday morning (10:15ish)
    • Andrew: Contact list about when we release the next major schema (and Paris)
    • Ola: respond to web start question
    • Brian: 2 QA tickets need loking at
  3. Status Update (20 mins)

    • Summary of last week visit (Will)
    • Namespaces for tags (Andrew)
  4. Testing Script (20 mins)

  5. AOCB (<5 mins)

    • ROI Mask additions?

Notes

  • Minutes accepted

    • Chris: annotation change seems fine
  • matters arising

    • Annotation
      • Changing schema, one rename
      • Chris: annotating annotations will be hard to code-generated
      • Code-gen went through previously
      • in schema and 4.2 releases
    • forum/email items were or will be handled
  • EM summary

    • reduction of submissions; not offering many features
    • looking into features which people want to bring them back
    • a certain number of papers require it (enforced?)
    • feature 1: more than the final structure but also the raw data
    • feature 2: maps/masks of chains (segmentation of volume)
    • OMERO: need image-image links
      • Use annotation as a proof-of-principle
    • Ingvar gave a presentation
    • writing MRC format?
    • next steps:
      • demonstrate interaction via OMERO
        • linking eman2 and spider and open astec viewer
        • they use all different languages / formats
        • will be a while before don't need external tools
        • focus on using processing as demo
        • deadline? * first year report in autumn
      • how to represent an EMAN2 project in OMERO (processing workflows)
        • currently uses a local berk. DB
    • Didn't discuss
      • communication channels
      • trac / tickets / etc.
    • Misc
      • Goal of using applet / viewer
      • Need for having a real user on board
        • Jason has already contacted some
        • Usual problem of selling the benefits
  • Namespace for tags

    • Andrew: bad form to not be filling the ns in our own annotations
    • Purpose of ns: gives us is to know where annotations came from
    • Chris/Josh: would rather have "ome.org/" rather than ".../"
    • Chris: we care about things that were imported & things exported
    • Also need to consider the relationship to Namespace
    • Jean-Marie: cF. mavkeMovie.py. Need to setup a standard
    • Josh: can use a ".../legacy" namespace for the upgrade
    • Jean-Marie: Currently use namespaces in a functional/interpretative way
    • Will: are we going to get collisions on the namespaces?
    • Chris: namespaces weren't designed for origins
    • Josh: namespaces were initially designed for defining the intpretation of the SA
    • Andrew: why did we just add "tag", it's another string?
      • "We think term/xml is good on their own"
    • Potentially use references to namespaces (top-level)
    • Agreed: have to clean up our own namespaces
    • Rest to be discussed later
    • Jason: @Tabling and asking in Paris?
      • Yes/no: do ask in Paris, but we need to get something ready for release
  • Testing scenario

    • Brian was supposed to prepare one (not present)
    • David had something on the UI website (couldn't be found)
    • @Will & Jean-Marie prepare something for Thursday
      • Writing up all the movies
      • Add that to a wiki page: http://trac.openmicroscopy.org.uk/omero/wiki/TestScripts
    • @Jason will provide a list in the next day and a half
    • http://wiki.secondlife.com/wiki/Category:Test_Scripts
  • Importer status

    • Jean-Marie tested Friday->Monday all test_images_good
    • 5-10% passed (imported at all)
    • Brian's changes as of Friday, ~15%
    • As of Monday with Chris and Brian, ~40%
    • As of now, ~75%

    • Goal by end of today is to have everything going in
    • Some are silly things (enum not matching, -60 for pos. integer)
    • Then will have to verify the metadata
    • Start on Thursday to do validation
    • Filters / filter sets going in (no light paths currently)
    • NB: Export image won't provide much information
    • 4.2 will export in entirety
    • Name accounts datasets with name of folder?
      • Using testengine to setup, and then need visual validation
      • Melissa? Y Carlos? N
      • Only test_images_good
      • See "New Metadata store" ticket (#2361)
  • Misc

    • Josh:
      • configuration / deployment code to be committed on return
      • database upgrade not done. may need to share plpgsql effort in June
    • Andrew:
      • RoiMask? @fix it!
  • Post-mtg mtg: annotations / namespaces

    • Donald: is ns "source" or "intent"?
    • Jean-Marie: for TermAnnotation we decided against version
    • Helpful to compare to XML?
    • Josh: NS/schema are extra layer of validation
    • Jean-Marie: But we don't have multiple namespaces like XML
    • Will: if you support multiple namespaces, then it's fine
    • Andrew: ns & origin discussion is not quite right
      • URL in XML NS about where it comes from
      • that part doesn't define what created the XML
      • "ome.org/tag" doesn't say "this tag from omero"
      • but "this interpretation of tag by omero"
      • Josh: can "" or null not implicitly mean "openmicroscopy.org"
      • Jean-Marie: making the model more usable
      • Andrew: fine with a default
      • Donald: trying to prevent empties
    • Using for 2 purposes
      • Rating namespace
      • Friday talked about using namespace term-annotation for external entity
    • Will: in some cases it's how to display in GUI (displayed, etc.)
    • Josh: is a single namespace just not useful enough?
    • Suggestion: drop the REQUIRED and suggest multiple namespaces in Dundee?
      • Andrew: still worried that we'll never get namespaces
      • Josh: could we not make empty ns's well-defined? (but some people will want to use them)
      • Donald: thought that the Namespace was the beginning of this
      • Josh: agreed, just need to test well, which is why it's available now
      • Donald: just merging the code today, starting to work with arrays of strings
        • good testing. should know. have good scripts for testing.
        • @Discuss in Paris
      • Jean-Marie: still think we mismatch source v. intent
      • Andrew: Caitlin had a notion of a creator (Photoshop)
      • Also discuss in Paris?
    • Josh: relationship to RDF
      • Donald: alternative storage?
      • Josh: still have to consider
    • Donald: should we in general ask about how people have extended OMERO
Document Actions