ABI Meeting minutes 2009-03-18

Last week's action items:

• Not yet done.  This action item to carry over. Action item 1)
  

• Gareth distributed the URLs.
• Randall asked how bugs will be found and fixed in the stylesheets/skin.
• Tommy said that Gareth is still working on the skin.

• Not yet done, as James is too busy to test it and Catherine is away.
• Randall asked if Mike could possibly test PMR2 Mercurial push access with some fake 1.1 models; Mike said that he can.
• Randall suggested Tommy make push access top priority.
• This action item to carry over. Action item 2)

This week's agenda:

• James wanted to get some discussion on this.

• This was deferred to a separate 11:30 meeting.

• Randall said that now that Alan is here, planning can continue.
• James' priorities are: COR-like text perspective and some work on the layout/interface.  He wants the existing functionality to be made easier to use.
• Peter suggested getting Alan and Justin to have a discussion of specific layout/interface changes.
• Justin said that it is now possible to execute arbitrary Javascript from the error console, and make live layout changes without recompiling. This could be adapted for use in supplying preset layouts.
• Andrew suggested that an extension approach might be better than scripting.
• Peter said that metadata editing is a very high priority and will require interface changes, this will need to be part of the interface discussion.
• Poul said that a consistent interface is very important for new users.
• Alan said that the COR-like text perspective is very high priority in order to get people using OpenCell for writing models; James agreed.
• Randall proposed a breakaway meeting on OpenCell planning.
• Randall mentioned tracker item 1624, KINSOL support.
• Andrew has been looking at IDA implementation.

• Alan reported on an Oxford project to improve on PyCml. Python bindings might be required.
• Randall suggested a breakaway for discussion of Python bindings.
• Alan said that people find the documentation and examples lacking for the Java bindings.  Alan was told by Jonathan Cooper that the CellML API is very comprehensive, but would benefit from 'proper' documentation.  We need better documentation and examples.
• Randall suggested that the API be documented for a particular language such as C or C++, then add some basic docs for other languages.
• Mike reported on attempts to build Java bindings with all options on his Windows machine; after a lot of effort and time he and Andrew have still not managed it.  There are a range of messy issues (API is currently built with very old versions of libraries) that require developer time to solve.
• Mike wondered if it would be possible to roll the Java bindings support for Windows with all options into the build process at every release.
• Andrew said that we do want to do automated builds for other platforms.
• Mike said that he also knows a group in Newcastle who want Java bindings.
• Justin agreed to roll the Java binding support for Windows into the build process using his own virtual machines.

• Mike asked if Justin could also build a Java wrapper with all options for Windows prior to 0.7 release, and Justin agreed to build  one for a stable snapshot.

• James has been thinking of writing some kind of very basic help docs for developers who want to use the CellML API.
• Randall said that what we are currently doing seems to work - that is, talking through the API with people. A good example is the productivity of Adam Reeve's summer studentship.
• Mike asked whether it was worth Andrew documenting what he shows new developers.
• Andrew said that it changes too fast to be worthwhile.
• Randall suggested that the website tell interested developers to "get in touch" as a minimum.

• Andrew has added code to the CellML API which means that generated code is always the same.

• Justin has finished multiple equation input.  There is not a snapshot of this yet.
• Andrew has been working on the text editor, which now scrolls much more smoothly.

• Tommy built the bulk of the skin for Gareth to work on.
• Mike would like a document on PMR2; concepts and jargon especially, to save time when he is testing.

• James has to set up the software centre for the new site.
• Randall suggested the possibility of a staged site for the workshop if the deadline can't be reached for a full launch.
• Poul would much prefer that the majority of the website is launched before the workshop.

• Tommy has been working on (a lot of things ?) code generation, setting curation tags, basic model searching, etc.
• Randall asked about authentication.
• Tommy said that users must be logged in to perform various actions, and that security policies can be defined.

• James has been working on the metadata spec, chopping it up and putting it up on github.

• James plans to put up Mike's models.
• Mike has a library of reusable components representing synthetic biology entities; models can be built from these parts.