Meeting Minutes 13 June 2007
Meeting Minutes 13 June 2007
Present: Randall Britten, Sarala Dissanayake, James Lawson, Tommy Yu, Poul Nielsen, Peter Hunter, Catherine Lloyd, Andrew Miller
Andrew noted that the bug tracker has been cleared of all major known issues. James is about 2/3 of the way through creating PCEnv functional tests, and should be finished by the end of today. Catherine has updated the tutorial in PCEnv but still needs to commit it.
There was some discussion about how Andrew should spend his time this week. It was decided that he would firstly look at whether there was anything he could do to try to improve MathML rendering very quickly, and if not, to turn it off completely for the release given that it is next to useless at the moment. After this he can look at helping Catherine and James with the documentation and testing of PCEnv.
Peter noted that there had been some dialogue by e-mail with Andre and Matt regarding the process for about the site, and Matt had suggested changes to how we handle changes to the website.
Andrew suggested that when Michael has finished site tools and Andrew Cantell has finished setting up the cellml.org website, it will become possible to automatically create test sites. Peter suggested that Randall and Tommy meet more regularly with Michael to check the progress of this and push along its implementation so we that the trunk can be shown on the development site for a while before changes get pushed onto the live site.
Andre was happy with the idea of having a final decision made in Auckland at the CellML meeting, but only if the issues were discussed first on the mailing list to allow all the issues to come out. Matt has proposed a more formal voting system. However, it was decided that such an approach is not really necessary, because as long as things are discussed on the mailing list first, everyones opinions can already be taken into account at the next meeting.
There has been a discussion on the mailing list about how the category system will work. It was decided that the current proposal, which will have a single level of categories (in effect, a one-level ontology) is good enough for now, but at some point in the future we may need to expand this to support a hierarchical system.
Tommy has been working on the keyword entry on the CellML site, and has also been looking at how to support sessions.
It was decided that Tommy will write up a proposal on how the repository can be re-structured to solve the issues with CellML 1.1 models and associating other files with CellML models. However, this is to take lower priority compared to the category issues.
It was agreed that the new system should assign random IDs to each model, but publications would be an alias to the randomly assigned IDs. Poul suggested that the models under the random IDs need not be versioned, and instead, the alias to the random IDs could be versioned (so for example, the main link for a peer-reviewed publication could be a link to the latest a random ID for a model, which is an unversioned directory. There would also be URLs which alias to older versions available). Poul suggested that Matt should also be involved in this process.
James has been creating PCEnv sessions. He has also been translating reactions into equations, and has been categorising CellML models.
He proposed the addition of several new categories, including 'viral dynamics,' 'immunology' and 'cancer.' He also had a plan to categorise electrophysiology models into cardiac electrophysiology and neuronal electrophysiology and an 'other electrophysiology' category. The group was unhappy with the idea of annotating models as other electrophysiology, as this would mean that models would have to be removed from this category if a new category was added. Andrew suggested that the 'electrophysiology' category be kept, so, for example, cardiac electrophysiology models would be categorised as both 'electrophysiology' and 'cardiac electrophysiology,' whereas models not fitting into an electrophysiology subgroup would simply be classified under 'electrophysiology.'
Randall suggested that categories be treated as 'saved searches', and this list be treated separately from the list of keywords. For example, there could be a category: Other electrophysiology model which was equivalent to an advanced search for "Electrophysiology and not cardiac", which cardiac electrophysiology would be "Electrophysiology and cardiac". Poul suggested that we also allow for custom saved searches. The consensus was that while this would be a nice feature, only the built in ones should be supported in the initial pass.
Andre said in an e-mail that he was concerned about the way we describe the model count, however, he would be happy if we made the words 'Peer reviewed' in the count a link to a page giving a definition of peer reviewed. Peter agreed to write up that page.
Sarala was experiencing difficulty trying to map CellML to BioPAX when there are imports of units, because the units are imported more than once. The group decided that Sarala should discuss it with Matt and Poul as the meeting was over time.
Andrew noted that the bug tracker has been cleared of all major known issues. James is about 2/3 of the way through creating PCEnv functional tests, and should be finished by the end of today. Catherine has updated the tutorial in PCEnv but still needs to commit it.
There was some discussion about how Andrew should spend his time this week. It was decided that he would firstly look at whether there was anything he could do to try to improve MathML rendering very quickly, and if not, to turn it off completely for the release given that it is next to useless at the moment. After this he can look at helping Catherine and James with the documentation and testing of PCEnv.
Peter noted that there had been some dialogue by e-mail with Andre and Matt regarding the process for about the site, and Matt had suggested changes to how we handle changes to the website.
Andrew suggested that when Michael has finished site tools and Andrew Cantell has finished setting up the cellml.org website, it will become possible to automatically create test sites. Peter suggested that Randall and Tommy meet more regularly with Michael to check the progress of this and push along its implementation so we that the trunk can be shown on the development site for a while before changes get pushed onto the live site.
Andre was happy with the idea of having a final decision made in Auckland at the CellML meeting, but only if the issues were discussed first on the mailing list to allow all the issues to come out. Matt has proposed a more formal voting system. However, it was decided that such an approach is not really necessary, because as long as things are discussed on the mailing list first, everyones opinions can already be taken into account at the next meeting.
There has been a discussion on the mailing list about how the category system will work. It was decided that the current proposal, which will have a single level of categories (in effect, a one-level ontology) is good enough for now, but at some point in the future we may need to expand this to support a hierarchical system.
Tommy has been working on the keyword entry on the CellML site, and has also been looking at how to support sessions.
It was decided that Tommy will write up a proposal on how the repository can be re-structured to solve the issues with CellML 1.1 models and associating other files with CellML models. However, this is to take lower priority compared to the category issues.
It was agreed that the new system should assign random IDs to each model, but publications would be an alias to the randomly assigned IDs. Poul suggested that the models under the random IDs need not be versioned, and instead, the alias to the random IDs could be versioned (so for example, the main link for a peer-reviewed publication could be a link to the latest a random ID for a model, which is an unversioned directory. There would also be URLs which alias to older versions available). Poul suggested that Matt should also be involved in this process.
James has been creating PCEnv sessions. He has also been translating reactions into equations, and has been categorising CellML models.
He proposed the addition of several new categories, including 'viral dynamics,' 'immunology' and 'cancer.' He also had a plan to categorise electrophysiology models into cardiac electrophysiology and neuronal electrophysiology and an 'other electrophysiology' category. The group was unhappy with the idea of annotating models as other electrophysiology, as this would mean that models would have to be removed from this category if a new category was added. Andrew suggested that the 'electrophysiology' category be kept, so, for example, cardiac electrophysiology models would be categorised as both 'electrophysiology' and 'cardiac electrophysiology,' whereas models not fitting into an electrophysiology subgroup would simply be classified under 'electrophysiology.'
Randall suggested that categories be treated as 'saved searches', and this list be treated separately from the list of keywords. For example, there could be a category: Other electrophysiology model which was equivalent to an advanced search for "Electrophysiology and not cardiac", which cardiac electrophysiology would be "Electrophysiology and cardiac". Poul suggested that we also allow for custom saved searches. The consensus was that while this would be a nice feature, only the built in ones should be supported in the initial pass.
Andre said in an e-mail that he was concerned about the way we describe the model count, however, he would be happy if we made the words 'Peer reviewed' in the count a link to a page giving a definition of peer reviewed. Peter agreed to write up that page.
Sarala was experiencing difficulty trying to map CellML to BioPAX when there are imports of units, because the units are imported more than once. The group decided that Sarala should discuss it with Matt and Poul as the meeting was over time.