Meeting minutes, 27th of June 2007

Present: Tommy Yu, Randall Britten, Catherine Lloyd, Poul Nielsen, James Lawson, Andrew Miller

James went through and summarised the mailing list items:

  • Curation status of models.

  • Wilfred Li noted that cellml.org was slow on the mailing list. Based on the data provided, it looks like the speed was not uncharacteristic of network speeds between here and North America, given that www.auckland.ac.nz took just as long to download. Poul suggested that perhaps we could consider a North American mirror if it is too slow. Randall suggested that perhaps Peter would be able to do a subjective test.

  • Andre has had an issue uploading models. Tommy has developed a workaround which will allow individual 1.1 models to be uploaded, although it still does not address the complete issue of putting up CellML 1.1 models.

  • Andre also noted that some models don't pass Jonathan Cooper's CellML validators. One option would be to have named flags instead of a simple starred system to reflect things like units inconsistency. Andrew suggested that this could be rendered on the page with additional icons, although James is unsure whether this will scale to 4-5 flags. James thinks that we should just put this sort of detail in a curation tab.

  • There was a discussion of whether this metadata should be put in the model. Andrew thinks that it should be kept separate because it can be added to individual revisions and we might not want to have to create a new revision to annotate an existing revision.

  • James is happy with the suggestion on the mailing list that he keep his curation process simple and follow the existing guidelines until there is a more detailed specification.

  • Tommy produced a document about the repository, and there is some discussion about it on-going.

  • There was also a discussion about the project team page. Andrew is opposed to the idea of having a project team page because it requires that we nominate a general CellML project team, which is detrimental to the idea of a community project. James noted that Matt had suggested that we justify the Auckland group as being the team because we have the initiative to get funding to work on CellML. However, Andrew noted that others like Alan Garny, members of Nigel Lovell's group, Jonathan Cooper, and various people at Osaka, Tokyo, and Kyoto universities also may have funding to work on CellML related projects. Poul thinks that they would then also have to be on the project team if we follow Matt's proposal. Andrew suggested that we avoid all the complexity of this and create a categorised page on which anyone doing anything CellML related can put their own projects and list the people. We could then have a category for projects such as curation of the cellml.org repository. Everyone at the meeting was happy with this proposal. James thinks that we should have a space for people to put information about what we are working on. Andrew noted that this could be achieved just be letting people use the Plone Members directory, as long as we create accounts for them. However, we still need a summary page with something more than a list of accounts.

  • The mailing list system at present is not working well. People at the institute such as Catherine and Jonna have unsubscribed or considered unsubscribing from the cellml-discussion mailing list because there is too much traffic which is not relevant to them. There was discussion about whether there could be any technological improvement, such as the use of a forum, to avoid this. Andrew is not convinced that this will do anything but move the problem, because people can use their mail client to get a threaded view anyway (or read the archives). Randall thinks that just because people can use their clients doesn't mean that they do. Andrew thought that one option might be a wiki, but Randall and Poul are not sure that would be any better. Andrew suggested that if we had a proper bug-tracker, such as Bugzilla, we could put issues in there, and have notifications of new threads sent to the mailing list. People who were interested could then comment through the bugtracker, and add themselves to the CC list for the bug. This discussion was deferred until other issues were resolved.

  • Andrew has created a proposed set of guidelines for the use of the team-cellml@list.bioeng.auckland.ac.nz. Everyone at the meeting agreed with the intent of the guidelines, but Poul thinks that the name of the list is inappropriate, and should be auckland@cellml.org. We could then allow other groups to use this facility and create lists if other groups want similar lists. These lists would all be publicly archived, and membership would be open to anyone. They will be exclusively for local messages. Andrew thinks that from a technical standpoint, it is an unusual design for people in distant locations to send mail to Auckland so they can communicate with each other. Poul thinks that this doesn't matter, and the advantage of this approach is that all CellML mail from around the world is archived and consolidated in one place, so the group agreed that we should do this. Andrew will write up a proposal about the change to send to the cellml-discussion list and solicit further feedback.

  • Andre has updated the graphing metadata specification.

Discussion then returned to the issue of the bug-tracker. Andrew thinks that Plone based bug-trackers like PloneCollector NG and Poi lack a lot of the features that we need. Andrew thinks that if we had a good bug-tracker like Bugzilla, we could use it for much of the discussions that currently occur on the mailing list. It would also mean that we could much more easily have one tracker to cover the entire CellML project or institute (at present we can do this with search, but Poi and PloneCollector NG aren't really very usable for this out of the box, so we end up with lots of small trackers, making it difficult to recategorise bugs, such as bugs reported by end users as PCEnv bugs that should be CellML API bugs and so on). The only problem brought up in the past with having a non-Plone tracker is that it doesn't integrate well with the Plone search. However, Andrew thinks that this is not a big enough issue to justify living with a bad tracker. Randall noted that the mailing list also doesn't integrate with Plone, and that isn't a big problem either. There was no objection to changing bug trackers at the meeting, so Poul suggested that Andrew write this up and get wider feedback. Andrew asked whether we should have a single tracker for the whole institute, or just for CellML. Poul thinks that if we can, having a single institute tracker is good. Andrew thinks that the problem with this is that again it makes CellML look like a non-distributed project, so perhaps we could have a single CellML specific tracker to cover everything from the CellML specification, through to any CellML software people are writing (whether Auckland or not), and any proposals that would currently be discussed on the mailing list. Due to lack of time this was not discussed fully.