2012-04-03 Tuesday Meeting
Attending: Andrew, Jean-Marie, Sebastian, Elwood (notes), Emma, Chris, Colin, Roger, Ola (excused 16:15), Will, Jason, Carlos, Liza, Gus, Scott, Stick
Remember: Agenda must be complete (with estimated times) on the day before the meeting. Any additions after that must go at the bottom (AOCB)
Housekeeping Pre-Meeting - 2pm Start
Matters Arising (<10 mins)
-
Actions
- Liza/Josh: update the ohloh logos: https://www.ohloh.net/p/bio-formats/contributors, https://www.ohloh.net/p/omero/contributors
- Everyone: again, please look for duplicates on ohloh: https://www.ohloh.net/p/omero/contributors, https://www.ohloh.net/p/bio-formats/contributors (Especially: Brian, Donald, Chris, Carlos)
- Andrew: improve handling of website feedback - done - now mails website-feedback list (andrew and emma)
- Ola: add webstart, downgrade, webadmin, basket to web discussion (see stories on milestone page)
-
Actions
Email and Forum Review - checked at (5 mins)
Email -
- Haari Jäälinoja - Re: [ome-users] Issues with NewUserGroupBean implementation (thread with Josh - last mail from Haari on 30 Mar - does it need a reply?)
Forums -
- Don't think there are any outstanding?
- Code forum review - none pending
QA -
link-- -- Add list indented 12 spaces ID Status User/email Type Created Selected file Comment Error 4266 new ian.dobbie@b... Insight 4.3.3 2012 04 02 MukB_D_A_ypet_07.dv Hi We are having problems uplo... java.lang.IllegalArgument... 4265 new fhauser@ucsd... Web 4.3.4-dbcb 2012 04 01 - still trying to fix Traceback (most recent ca... 4264 new hsjk1@ku.edu... Importer 4.3.3-00d1 2012 03 30 ...age Display-Acquire CCD -0.tif Images were exported from FEI'... java.io.IOException: 1: b...
Agenda - 2:30pm Start
Accepting minutes from last meeting
Sprint Bug Report (5 mins)
- New bugs: 12 (6 critical, 1 major, 5 minor)
- Total bugs: 97 (4 blocker, 12 critical, 29 major, 52 minor)
Colin: FS presentation (45 mins)
4.4 Decisions (15 mins)
Testing (10 mins)
Satellite Update
Glencoe Update
Any other business (<5 mins)
Notes
Attending: Andrew, Jean-Marie, Sebastian, Elwood (notes), Emma, Chris, Colin, Roger, Ola (excused 16:15), Will, Jason, Carlos, Liza, Gus, Scott, Stick
Pre-Meeting All addressed, repsonded and covered
Agenda - 2:30pm Start
Accepting minutes from last meeting
- None
Sprint Bug Report (5 mins)
New bugs: 12 (6 critical, 1 major, 5 minor)
Total bugs: 97 (4 blocker, 12 critical, 29 major, 52 minor)
- Emma: Total 97, close to 100
-
Jean-Marie: testing branches effort, will go down after testing the web, 10 new ones from Gus and Scott in Insight, as soon as those branches go in they will go down
- Talked to Will & Josh that josh had chmod branch and we merged it in to develop. Tests written on it, so Will Carlos and I can start work. So in place for when Josh comes back
-
Ola: Email for an import issue using FS
- Chris: Size of file an issue?
- JM: using FS
- Ola: no imports found
- Chris: up logigng to debug
- Ola: tried set of images and a single image
- Colin: logging on cls-dev machine
-
Chris: trtack exception, probelm with those files is to short circuit the min-max calculation
- Trying to calc min max on a large jpeg, decompressing
- Colin: If not a big image is done inline
Colin: FS presentation (45 mins) - Start 14:33
- fs-lite and import
- first few slide repeated from July, old import process and looking at new one
- all in ticket 909
- numbered files are not useful / helpful are database related
- import libraries on the client
-
big images added a bout a year ago
- we always archive the big image
- after import a separate process worked through the pyramid generation on big images
- lots of duplication/ triplication of data
- templated directory structures
- on import tries to work out a unique path
- new import process gets the unique path and then uploaded to the server
- control handed over to the server processing of uploaded file
- dropbox was duplication but under new scheme it doesnt
- down side is we have to access the pixel data from the original file, some overhead to work out for different file formats
- no performance testing yet, in progress
- file path clashing issues, common file names, e.g. log files, summary files
- orignal metadata needs to be generated on import
- notifications
- cancellation after all files have been uploaded
- discussion over exposing file system
-
Roger: read-only after import?
- Chris/Colin: no they are writable would be an admin concern
-
Jason: file system level i deleted everything in 2012, in pricipal it is recreatable if i have original data.
-
Chris: only problem is we dont summaries import effectively in the database. Wouldnt be able to tell which file created which image.
- LEI stores summary file, to recreate the information it would require a lot of work
-
Chris: only problem is we dont summaries import effectively in the database. Wouldnt be able to tell which file created which image.
-
Jason: Archiving?
- Chris: get it for free
Done 15:30
4.4 Decisions (15 mins)
Jason: 4.4 release of fs?
- Chris: do or dont, no inbetween
- Colin: perf. testing this week
- Jason: CRUK data?
- Colin: cls image repository testing
- Chris: HCS use case testing
- Emma: wont make 4.4 but some time after.
- Chris: keep Colins branch going but release to a subset of installations
- Chris: impacts ALL users
-
Colin: preview versions?
- Marking importer as deprecated, importer less deprecated than stuff that doesnt work
- Chris: do we release this to the wild? or do we approach the proper use case, showing them the file system by default?
- Jason: trials at 10 or so places
- Chris: all imports are on the server
Emma: meeting on Friday
- Jean-Marie: some decisions pending, channel/logical channel, rendering (4.4.1)
-
Jean-Marie: FLIM floating point -> int values?
- Jason: yes
-
Jean-Marie: large pixel ranges, image comes in but still black
- Chris: we dont support any more than 16bit of dynamic range
- J-M: min max range is bigger
- Jean-marie: Insight headless, made good progress, from Dresden. For multi-thread need to do some cleaning. hopefully in 4.4 maybe, 4.4.1 definitely
- Chris: Rendering Engine ones not in 4.4
-
Chris: Channel/Logical Channel we need to decide collectively, impacts everyone!
- Jean-marie: The other model changes, naming conventions have been made.
-
Will: Channel/Logical channel merge is less imapct?
- Chris: equivalent impact, properties move, layer of heirarchy moves, queries updated, objects removed from mapping level, not trivial
- Jean-Marie: omero matlab, omero py, folloing ian's email to allow for them to look at it. For name change, get channel -> get logical channel
- Chris: only place in entire model where we have inconsistency if we don't do it in structure
- Will: saving breaking change for after 4.4
- Jason: amount of time to accomodate that, additional time to do logical channel? sprint or 7?
- Jason: project wide sprint,
- Chris: all docs, clients, servers, insight, insight gateway
- Will: no exposure in blitz gateway, there is documentation though
- Chris: wrapper objects written by hand, all needs to be updated, scripts
- Jason: extended by 2 weeks? The model is correct!
- Will: seeing as were doing it as a breaking change release.
- Chris: whole bunch of havoc by removing many-one, in omero you can have multiple pixels, not prevented.
- Jean-Marie: extend by another sprint.
- Elwood: Testing after the work has been done.
- Emma: Decision made, do it and extend by 2 weeks.
-
Jason: Code generation / database?
- Chris: Code-gen done
- Chris: Database changes pending
- Andrew: Changes are detailed on tickets 8084,5,6
- Jason: OME calendar, schedule between now and Paris
- Final model changes
Testing (10 mins) - 15:50
- Emma: testing being generated. Need thought and discussion on who and what will be allocated. Come next Tuesday / after, expectations are understood.
- Emma: Next Tuesday people are away so schedule for later this week.
-
Jason: Get some prep done before we meet
- Making it available and what should be tested
- Chris: localised changes we've been doing okay. ON database changes it effectively test everything, several weeks form that, at least 2 weeks from that if not 4-5 from being able to start that.
- Jason: between now and then is there the possibilty to test what is in place now?
- Chris: coverage on file formats is good from Roger
- Chris: whole scale testing apart from small stuff is Doms UI in web client. Nearest wholescale testing operation.
- Jason: need to address issue with Scott and Gus how soon do we hand off areas, next full on testing is web UI
- Chris: as soon as the UI is functionally complete, need to touch whole interface elements.
- Will: plugin branch has some wide spanning changes too
- Jason: articulate what is needed in terms of testing the UI.
- Chris: Probably not everyone on the entire team.
- Ola: across all web browsers, the new UI
- Gus: if testing is massive, is it doable in the timescale?
- Gus: big backlog for user support team already. I browse and fiddle to find problems.
-
Jason:
- need a schedule for when the web UI is going to be available, within a week.
- Then we need a schedule for when model will be changed and who will be available.
- 3.1 If UI testing how can people get access
-
3.2 What the process will be when we start testing?
- Testing scenarios has been a bit too atomic before.
- Chris: localised testing has been more thorough.
- Jason: waste of time to repeat those, what is the right way to test the application before release?
- Jean-Marie: what was quite efficient was when we did movies with a real workflow.
-
Chris: you will come across workflow related issues, where we accept that its the way it works
- Advantage is we're doing it soon enough to potentially resolve some of those
-
Jean-Marie: Use movie list as start of test plan.
- Scott: we need enough details for someone to follow.
- Gus: needs context when testing to better find and follow expected workflow.
Satellite Update - 16:15
- Mailing list checklist
- Oxford short list candidates
Glencoe Update - 16:17
- Big images - scale and quantity
- Server requirements for performance needs documenting
Any other business (<5 mins)
Chris: hudson problems with builds, working more with it. Speaking to stick for next web testing
Actions * Jean-Marie: Talk to melissa about a QA issues (4264, 4266) * Chris: Talk to stick about web testing infrastructure.