Meeting Minutes 25 July 2007

== Meeting minutes, 25th of July 2007 == Present: Sarala Dissanayake, Catherine Lloyd, Randall Britten, James Lawson, Poul Nielsen, Andrew Miller (speakerphone,) Matt Halstead, Peter Hunter
James and Randall suggested we re-evaluate the importance of James summarising the traffic on the discussion lists each week. It was decided that a summary was not required - people can either read the list or the archives for this purpose. If there is an issue on the lists that needs to be discussed or have a decision made on at the Auckland CellML meeting, this should be sent to Randall as an agenda item. Randall suggested that perhaps we set up an agenda page/wiki on the website to facilitate this.

Matt brought up the issue of threads that start in the team-cellml list but should be moved to the cellml-discussion list. He suggested that we assign a person to be responsible for summarising the key points discussed in the thread and then posting a coherent message to the cellml-discussion list, rather than simply forwarding/moving the thread, which can often be jumbled and confusing for people who have not been following the original thread in the team-cellml list. It was agreed that this was important, and the group was happy with taking shared responsibility for this. For example, if the thread is about PCEnv, then Andrew Miller might be responsible for this, as he would have been following all the posts in this thread.

We discussed the workflow of issues in the mailing lists being raised to tracker status. The main concern is that many people will not read the tracker daily, but will follow threads in a mail list, so that threads that continue in tracker issues will become invisible to those only reading the mail lists. Matt suggested that at first a discussion about an issue should reach a point on the discussion list where anyone who felt inclined to has sent their 2 cents worth. Then to promote the issue to the tracker and indicate to the list that people should join the thread of that issue if they had any further interest in it.

The proposal to create a new email list with the address "" was discussed. This was raised several meetings ago but postponed until Peter arrived back from overseas. The idea was to create a membership-moderated, open-archive list specifically for the Auckland CellML team to use. In previous meetings it was suggested that other groups could also get their own '' list. Poul would prefer not to create an '' list, as he questions what it would be used for. He favours keeping the team-cellml list in use as it is, but changing the domain from to, in keeping with the idea that this is really a community list, not an Auckland-specific list. There was general agreement on this issue. We will talk to the IT staff about how this can be done. The team-cellml list is currently private, and the group agreed that the new list should have publicly available archives. In general, the list should be open for people to join at request, but Peter is keen to not let just anyone in to give their tuppence. It was suggested that requests for membership should go through Randall, who would then bring the requests to the Auckland meeting. Ideally, the more regularly contributing members we have on this core list, the better. If the archives are to be public, members should take care when discussing potentially sensitive subjects, such as other peoples'/groups' work.

Peter gave a run-down on the decisions that have been made over the last few days re: tracker selection, evaluation and implementation. It has been decided that we will use Bugzilla as an interim solution. Until such time as we have evaluated the options and selected a new tracker, people should use Bugzilla. In the mean time, we should evaluate Poi/Plone Software Centre, Bugzilla and Trac. Randall has drawn up a list of the key requirements we have for a tracker, which he will discuss with Matt. Matt is happy to set up several different tracker products for evaluation, and will mirror bugs/issues between the trackers. Ideally, people should try to get a feel for all the trackers under evaluation by submitting each issue to all of them. If people do not wish to do this, then bugs/issues should go into the Bugzilla tracker, which is currently at:

The current decision with the bug tracker is:
  • that we do not publicize any particular tacker as the cellml tracker just yet.
  • we set up trac (in a couple of guises) in addition to the current bugzilla, poi+plone software center scenario to offer three different systems to enter issues into
  • over the next 3-4 weeks, the core group of developers, modellers, and cellml team members can triple submit their issues - i.e. add their issue to each tracker - and then update their own notes on what they liked and/or didn't like about the various options
  • bugzilla's database of issues will be treated nicely and backed up properly so that at the end of 3-4 weeks we have a single source of issues that we know contains all the valuable data that has been added. The others will be backed up too, but we must keep checking the integrity of at least the bugzilla database
  • at the end of 3-4 weeks we will make a decision about what technology we want to use and deploy it immediately ourselves independent of the IT group and advertise it to the public.
  • longer term we will discuss the deployment issues within the ABI and what resources are needed
  • this process is complicated by the fact that Peter would like to see the same software stack used to resolve this tracker problem also used for CMISS, CMGUI. We don't anticipate replicating the test described above across those projects which have active public participants, but want to know how to include important reviewers of technology (like Shane and Andre) in to this CellML test with the incentive to project what we learn here on to those other projects.
There was then a discussion about Sarala's work and annotation of models with biological information. The group discussed whether James should be prioritising bread-and-butter curation or whether he should be putting time into more theoretical work such as looking at how models can be broken down into components and rebuilding models using 1.1 architectures. Matt is going to work on a method based on the bioentity metadata for James and Catherine to start annotating models with biological information in such a manner that a script can be used to render this information into a format that Sarala can use.