2007-11-15 Project-wide 3:45 GMT
Attending: Alexa, Andrew, Brian, Chris, Curtis, Jason, JMarie, Josh, Kevin, Will
Agenda
- OME-XML
- validator
- xsd-fu (working with OME-XML in Java)
- New API requirements from Curtis?
- Potential usage inside OME Notes and the Protocol Editor
- OMERO 2.2.1/2.3
- Permanent generation bug (#834)
- Rendering engine de-synchronization and memory leak
- Searching
- use cases wished
- easy: search("Image", "name:*bob* OR description:...")
- tough: search("ImageName:*bob* AND ImageTag"
- questions:
- default field? implies that the user doesn't need to add "field:..." in query
- building a UI? need metadata for client
- use cases wished
- Moving OMERO towards OME Server functionality
- use cases written over Thanksgiving
- also requests from CMU
Action Items
Notes
- ome-xml
- waiting for people to complain
- records on usage?
- anything from improvision?
- xsd-fu
- ctor(parent node object)
- typed?
- currently only works from one schema
- back-references (not urgent for new schema)
- #sampleTest isn't complete
- curtis working on generation
- some methods missing from ome-java
- use inside ome-notes/protocol editor
- more methods to ask questions about hierarchy would be good
- simplifies use of working with XML
- certain points in hierarchy: "what fits here below instrument?"
- also a goal of generating XML scratch from any starting point
- should have a discussion on notes/protocol editor synergies
- acquisition metadata vs. experimental metadata
- for example: can you call bioformats during import to add parameters
- ctor(parent node object)
- cellbio
- kevin ordered all booth stuff
- has ilya ordered electrical?
- OMERO 2.2.1 memory issues
- valkyrie happy with new runline options
- 95%-99.99%, non-monotonic
- should we fiddle more?
- increase heap size?
- should be able to test new RE on Monday
- moving OMERO towards OME server functionality
- jason will write up, but had to run
- searching
- jmarie started working on UI for searching
- will contact josh friday morning
- "any term that comes into your head"
- two dimensions: field and class
- you might not know the type
- know difference between tag and a description field, e.g.
- within transactional boundaries
- owner information
- can everything be in one index
- prohibitive
- searching between indices
- have to join at build time
- can find a name somewhere
- putting everything in one index
- (XML) content: with everything in it
- also: owner, group, type, tags, files(?),
- planning
- need to be able to tear it down
- flexibility
- need to reduce query-space as soon as possible
- quotas
- column in the password table?
- 2.2.1
- possibly another couple of formats
- but we shouldn't drag it out very long