2016-04-26 Tuesday Team Meeting
Dundee: Petr, Will, Simon, Mark C, Jean-Marie, Gus, Dominik, Colin, Simone, Ola, June, Kenny, Balaji, Helen, Jason, Roger Remote: Ian, Josh M, Melissa, Sebastien B, Kelli, Liza, Andrea, Chris, Eleanor, Stick, Wilma
Agenda - 2:30pm Start
Accepting minutes from last meeting
- accepted
Project Timelines (2-3 minutes each)
Spaces - 14:31 UK
-
Mainline (J-m)
- OMERO PR freeze now in effect
- further testing with Ice 3.6 this morning, working fine
-
RC later this week to test changes with downloads page, etc.
- if testing goes well, full release early next week
-
Model (Sebastien)
-
reviewed board to check last breaking changes
- will need to discuss a couple of cards with David
- ensure OMERO can consume them
- cleaning up Insight after issues from demo
-
created new OMERO 5.3.x board in addition to the 5.3.0 board
- "buckets" become columns
-
after coming OMERO release scoping should begin in earnest
- may report progress at these meetings
-
reviewed board to check last breaking changes
-
Metadata (Josh M)
- Tagged IDR 0.0.4 build (OMERO 5.2), deployed on a couple of servers
-
Beginning re-import of mitocheck as well as testing of new datasets.
- docker, openstack
- Testing new caching configuration and preparing for web-proxy migration
- No pressing need known for IDR to be on OMERO 5.3
-
Mainline (J-m)
Other releases/upgrades - 14:39 UK
-
Figure
- Will: minor release last week
- Jean-Marie: planning/scoping for further work? (ROI, tag...)
- Jason: put them in an inbox column to make them all explicit
- ImageJ
- Learning
-
Sysadmin
- Kenny: nginx upgrade on haddock and nightshade, trying to increase consistency
- Kenny: EasyDNS renewal done
- Josh M: recent reliability issues with docker, suggests move to OpenStack
-
Figure
Glencoe Update (Chris) - 14:44 UK
-
working with Harald on unit / marshal issues, making progress
- also looking at masks
- hoping to share ROI code soon
-
working with Harald on unit / marshal issues, making progress
The Great Demise of Jabber (Jason) - 14:45 UK
- considering switching to Slack, speak up soon if you have a strong opinion
OME Mtg Update (Jason & June) - 14:49 UK
- June: 55 people registered (32 externals), closes this Friday so needs a last push to remind people to register
- Jean-Marie: could remind local people, e.g., Ninewells
Social media update (Kelli) - 14:52 UK
-
Kelli in US over May, back in June
- Facebook, LinkedIn on hiatus
- Helen to do announcements, Josh handling interesting links, etc.
-
Kelli in US over May, back in June
nDimension presentation by Roger - 14:54 UK
- Mark: allowing open naming of dimensions raises community issue of same people using same dimension name for different things
- Roger: imglib2 may be interesting to study in how they handle storage
-
Ian: requesting pixels in not their storage order could mean reading the whole file many times
- Roger: yes, storage order can make a big difference, though that order is available to caller so they can optimize accordingly; also region copies can be optimized by the library where a set of memcpy could be done
-
Simon: could it help with acquisition in allowing microscope to stream in any order?
- Roger: is already flexible by IFD layout in TIFF
-
Will: could allow sparse arrays?
- Roger: possibly, as an implementation detail
- Roger: for higher-dimensional data sparseness may more probably arise (samples aren't over exactly a hyperrectangle)
-
Jason: helps correlative / micro-magellan / non-contiguous sampling
- Josh M: not really
-
Jason: avian? (expose Java to C++) prefer not to reimplement imglib2
- Roger: probably wouldn't allow low/zero-copy, etc., so performance would suffer; would be useful for wrapping existing readers
- Roger: imglib2 would probably be worth looking at for a Java implementation
- Jason: check out vigra
-
Roger: still need to do the model part anyway
- Josh: good to compare with others' nDim approaches
- Roger: would be pleasant surprise to find any offering PixelBuffer for patterning data for consumption by OpenGL, etc., etc.
- Roger: proposed C++ code generalizes Boost multi-array library
AOB (5 mins max - technical discussions should be highlighted to relevant people and rescheduled) - 15:45 UK
-
CTIR Scientific Advisory Boards (Dundee only)
- SAB visit on May 9th
- open series of talks by PIs
- opportunity for Dundee OME team to meet Søren Brunak, our SAB member
- not mandatory but useful and encouraged
-
CTIR Scientific Advisory Boards (Dundee only)
Done 15:53 UK