Meeting Minutes 16 May 2007
Meeting Minutes 16 May 2007
Present: Poul Nielsen, Tommy Yu, Catherine Lloyd, Jonna Terklidsen, Randall Britten, Dave Brooks, Sarala Dissanayake, Andrew Miller
James has been working on addressing the models which have reaction components. Andrew wrote a script to help him address this. The script works by moving mathematics out of the reaction element and into the containing component, and also by adding new equations describing the relationship between the 'delta variable' and the rate equation based on the type and direction of the reaction, and the stoichiometry specified in the role.
Andrew noted that his script didn't automatically remove the reaction element, because this contains information which needs to be preserved somewhere. Poul suggested that Matt should have input into this process.
Sarala has been generating OWL and BioPAX from CellML models, so that this can be visualised. She is not currently dealing with the reaction element. Poul thinks that this is the right decision, because we want to move reaction in the metadata and deprecate reaction.
Andrew has now disabled the integrators which we decided weren't worth having in PCEnv, by commenting them out in the PCEnv front end. They are still available through the CellML Integration Service programmatic interface.
Andrew has now finished the implementation and testsuites of MaLaES (MathML to Language Expression Service), CeVAS (CellML Variable Association Service) and CUSES (CellML Units Simplification and Expansion Service). However, there is still a need to make CCGS make use of these services. The benefit of this is that CCGS can be re-engineered to use MaLaES and support a range of different languages, as well as a better model which allows for derivatives to appear anywhere in the equations. The downside of this is that it will take away development time from the fairly long list of PCEnv usability issues, and it will require a CCGS interface change. The consensus was that I should work on the CCGS refactoring first.
Dave has been looking at how FieldML can be implemented. He has been looking at Geographic Markup Language (GML) because it already has an underlying model which allows for arbitrary fields to be defined. He will look at how this and other specifications could be put together to create models in a similar way to how MathML can be leveraged into CellML. He noted that GML has already been through a standardisation process and become an ISO standard.
Randell has been talking to a range of people to try to learn about CellML and other technologies being used.
Tommy has been working on properly supporting variants. He has also improved the modification history from the previous XSL based view. He noted that there are still issues with RDF corruption due to 4Suite bugs and bugs in the old CellML repository. He has written a script to fix some of the issues, but when corrupted models have been read into other software and then re-saved, it is too hard to recover them, so these models have to be fixed by hand starting from the model in the CVS used to generate the models. James might be able to help in this process, because he has a log of the changes that he has made to the metadata. Poul is concerned about what process we can use so that we are not losing important curation metadata. Andrew thinks that in the future, this could be addressed just by having regular backups of the repository, and testing code against the current repository using Michael's site-tools when it is ready.
James has been working on addressing the models which have reaction components. Andrew wrote a script to help him address this. The script works by moving mathematics out of the reaction element and into the containing component, and also by adding new equations describing the relationship between the 'delta variable' and the rate equation based on the type and direction of the reaction, and the stoichiometry specified in the role.
Andrew noted that his script didn't automatically remove the reaction element, because this contains information which needs to be preserved somewhere. Poul suggested that Matt should have input into this process.
Sarala has been generating OWL and BioPAX from CellML models, so that this can be visualised. She is not currently dealing with the reaction element. Poul thinks that this is the right decision, because we want to move reaction in the metadata and deprecate reaction.
Andrew has now disabled the integrators which we decided weren't worth having in PCEnv, by commenting them out in the PCEnv front end. They are still available through the CellML Integration Service programmatic interface.
Andrew has now finished the implementation and testsuites of MaLaES (MathML to Language Expression Service), CeVAS (CellML Variable Association Service) and CUSES (CellML Units Simplification and Expansion Service). However, there is still a need to make CCGS make use of these services. The benefit of this is that CCGS can be re-engineered to use MaLaES and support a range of different languages, as well as a better model which allows for derivatives to appear anywhere in the equations. The downside of this is that it will take away development time from the fairly long list of PCEnv usability issues, and it will require a CCGS interface change. The consensus was that I should work on the CCGS refactoring first.
Dave has been looking at how FieldML can be implemented. He has been looking at Geographic Markup Language (GML) because it already has an underlying model which allows for arbitrary fields to be defined. He will look at how this and other specifications could be put together to create models in a similar way to how MathML can be leveraged into CellML. He noted that GML has already been through a standardisation process and become an ISO standard.
Randell has been talking to a range of people to try to learn about CellML and other technologies being used.
Tommy has been working on properly supporting variants. He has also improved the modification history from the previous XSL based view. He noted that there are still issues with RDF corruption due to 4Suite bugs and bugs in the old CellML repository. He has written a script to fix some of the issues, but when corrupted models have been read into other software and then re-saved, it is too hard to recover them, so these models have to be fixed by hand starting from the model in the CVS used to generate the models. James might be able to help in this process, because he has a log of the changes that he has made to the metadata. Poul is concerned about what process we can use so that we are not losing important curation metadata. Andrew thinks that in the future, this could be addressed just by having regular backups of the repository, and testing code against the current repository using Michael's site-tools when it is ready.