Personal tools
  • We're Hiring!

You are here: Home Community Minutes Conference Calls 2007 2007-10-04 Project-wide 3:45 BST

2007-10-04 Project-wide 3:45 BST

Attending: Alexa, Andrew, Brian, Cat, Chris, Curtis, Donald, JMarie, Josh

Agenda


  • Curtis
    • OME-XML Node stuff (Chris)
      • need additional node types? want additional methods?
      • getting parent node — e.g., get Filter for a given TransmittanceRange
      • getting nodes that reference this one — e.g., get Images that reference a given Detector
      • methods to add and remove children — e.g., add DetectorRefs to an Image
    • Metadata revisioning (Josh)
    • OME-XML/OME-TIFF timeline
  • Chris
    • RE exception
  • Andrew
    • XML goodies


Notes


  • curtis
    • prototyping node types for new version of XML
      • previous version only looked at STD
      • now coming directly from generateDS (python)
      • do you (chris) want more node types? or do you want more?
      • Chris: currently generating for every element in xsd (almost compiling)
        • few methods needed in OmeXmlNode
        • few tricky things defined in global scope & used only once (~10)
      • Curtis: Would like the hierarchy to be "true" (extensions, etc.)
      • Andrew: ExperimenterType disappeared
      • Curtis: great. solved that weirdness
      • Curtis: test is not complete, but notes on everything
      • Chris: No set Node
      • Curtis: use Child(Parent) ctor
      • Curtis: Should make sure "all the nodes that reference me" still works in 2007
    • metadata versioning
      • not so much about SVN
      • repeated
        • image in db
        • check it out (and metadata)
        • edit it
        • commit back (but not delete)
      • copies and deletes
        • client needs to be smart
        • doesn't undo globally
        • want to be able to ignore others people
      • tricky part is the merge
      • shouldn't be too hard to add a "branch" tag to the metadata
      • discuss at the next face-to-face
      • Curtis will start a list of use cases
    • OME-TIFF
      • once the mapping work (see above) is done, things should move forward
      • Curtis: cool long term solution, but the short term solution?
        • Using loop back in bio-formats. Why not use that?
        • Chris: matter of prioritization, how much time for what? (papers/doing it correctly)
        • Curtis: quick way should only be few lines of code. Then we support OME-XML.
        • Andrew: Anything to be done on XML to help this along?
        • Andrew: Generating validly-wrong files!
        • Curtis: keeps Andrew posted on dev. of sample data generator
        • Chris: An authoritative list would be nice (for OMERO, validator, .... )
          • Consistent ERROR numbers across all applications
          • only have to write help for the error list once
          • Andrew will start that list up on ome-xml.org trac
    • Multiple file name-law
      • stitching logic...
      • Scan all in the same folder, and check for ids in the files?
      • Something in the XML to describe the other files?
      • Doesn't solve renames; changes the schema (optional)
      • .ome.tiff as proper extension?
      • Add to all sample datasets for parser-writers
      • Official proposal (for next non-breaking release): Add optional "Path" attribute to TiffData with pointers to all other files
        • Andrew: next one is breaking
        • Could also allow for sub-folders
        • Allow absolute paths?
        • What about the later files? ... Already dealing with XML-block generation built over file and inserted at the end.
        • What time for slashes? Forward-slash, good in Java. Need to qualify. Are c:\ allowed?
  • Andrew
    • Other than the above, have only added fixing PixelsType to todo list
Document Actions