Meeting Minutes 6th November 2006
Meeting Minutes 6th November 2006
Present: Sarala Dissanayake, Poul Nielsen, Andrew Miller, Shane Blackett, Jonna Terkildsen
Poul noted that there had been discussions on the mailing list, relating to the release of the CellML API, XPCORBA, xpcellml_api and PCEnv.
Shane was keen for more than one program using the API. Currently PCEnv, as well as Andre's program, do use it, but Shane is concerned they may not be sufficiently independent in their use of the CellML API, as both use the CellML Code Generation Service.
Jonna noted that she found it hard to find variables in PCEnv, because there were too many in complex models. Viewing them by components would help, as would being able to sort variables. She would also find it useful if there was a way to specify a set of variables which she wants to see, and hide all the other.
Poul suggested that we could use the grouping facilities. However, there are two issues with this:
1) CellML groups apply to components, but it is currently not possible in CellML to group variables.
2) The CellML specification has, so far, tried to separate data, which affects the interpretation of the model, from metadata, which, controls how the model is displayed. It seems that this is clearly metadata, and so it should be in the RDF.
Andrew will put out a draft specification which uses RDF to define groups of variables which can be displayed, and post a link to the draft on the mailing list.
Andrew noted that the specification for the CellML media type has been published as an RFC (RFC4708). He asked where we should add information about this (Poul suggested a new section of the website for this). Andrew also noted that the CellML repository still seemed to be serving models as text/xml, which Andrew will look into (also need to inform cellml-discussion, to warn them of the change).
Andrew asked how long we should hold off releasing the CellML API 1.0 after the release candidate (if no problems are explicitly identified). Poul thinks that we need to leave it for a reasonable length of time, to allow people to try to implement this. Andrew is concerned that we could end up maintaining too many branches. Shane thinks that we should maintain the branches even after the 1.0 release, rather than requiring people to move on to another release to get the latest bugfixes.
Poul noted the discussions about changing the format of the meeting, and noted that he was not convinced that we should stop having regular face-to-face meetings. Andrew noted that, as discussed last week, he had looked into whether Jabber has the features required (so that it can save previous discussion, like Campfire does), and mu-conference appears to do this. He has set up a test Jabber server on his system (and posted details to team-cellml) so people can test this out.
Poul noted that there had been discussions on the mailing list, relating to the release of the CellML API, XPCORBA, xpcellml_api and PCEnv.
Shane was keen for more than one program using the API. Currently PCEnv, as well as Andre's program, do use it, but Shane is concerned they may not be sufficiently independent in their use of the CellML API, as both use the CellML Code Generation Service.
Jonna noted that she found it hard to find variables in PCEnv, because there were too many in complex models. Viewing them by components would help, as would being able to sort variables. She would also find it useful if there was a way to specify a set of variables which she wants to see, and hide all the other.
Poul suggested that we could use the grouping facilities. However, there are two issues with this:
1) CellML groups apply to components, but it is currently not possible in CellML to group variables.
2) The CellML specification has, so far, tried to separate data, which affects the interpretation of the model, from metadata, which, controls how the model is displayed. It seems that this is clearly metadata, and so it should be in the RDF.
Andrew will put out a draft specification which uses RDF to define groups of variables which can be displayed, and post a link to the draft on the mailing list.
Andrew noted that the specification for the CellML media type has been published as an RFC (RFC4708). He asked where we should add information about this (Poul suggested a new section of the website for this). Andrew also noted that the CellML repository still seemed to be serving models as text/xml, which Andrew will look into (also need to inform cellml-discussion, to warn them of the change).
Andrew asked how long we should hold off releasing the CellML API 1.0 after the release candidate (if no problems are explicitly identified). Poul thinks that we need to leave it for a reasonable length of time, to allow people to try to implement this. Andrew is concerned that we could end up maintaining too many branches. Shane thinks that we should maintain the branches even after the 1.0 release, rather than requiring people to move on to another release to get the latest bugfixes.
Poul noted the discussions about changing the format of the meeting, and noted that he was not convinced that we should stop having regular face-to-face meetings. Andrew noted that, as discussed last week, he had looked into whether Jabber has the features required (so that it can save previous discussion, like Campfire does), and mu-conference appears to do this. He has set up a test Jabber server on his system (and posted details to team-cellml) so people can test this out.