On 7/18/07, David Nickerson <[EMAIL PROTECTED]> wrote: > Hi Matt, > > > I think there does need to be some description of what the parts of a > > cellml specification version number is and how it relates to the > > cellml namespace(s). > > > > My understanding is that the 1.1.1 was chosen so that this did > > represent a minor step in thinking for CellML 1.1, but that the model > > remains valid as a 1.1 model. i.e. A CellML 1.1 model is not > > necessarily a valid 1.1.1 model and a migration step would be needed > > to make it so. The namespace of a 1.1.1 model would still be > > http://www.cellml.org/cellml/1.1# > > > > I also think that this minor version should be marked in glaring red > > that it is precisely a depreciation of reactions step and provide > > reasons why and the tools to help people transform any models they > > have with reactions into ones that don't. I think there will be some > > debate around that - especially pointing to lossless transforms > > between SBML and CellML and back again with existing transforms > > available out there. > > > > My assumption would be that CellML 1.2 or 2.0 namespace would be the > > one where reaction elements become invalid if indeed 1.1.1 was > > successful. > > you seem to be saying that version 1.1.1 is marking the reaction element > for deprecation but that the reaction element itself won't become > invalid until a 1.2 or 2.0 version is released? This doesn't seem to > match with completely removing all reference of and to reaction elements > in the 1.1.1 specification...maybe I'm missing something? >
No. I agree that removing reference to reaction elements as they exist in 1.1 is probably confusing and I think 1.1.1 (or whatever it will end up being called) should emphasise that it is this depreciation of reaction elements that the 1.0 and 1.1 specifications contain. But we do need to have a resulting specification (including its XML DTD and Schema) that has these reaction elements removed. In essence a specification that is devoid of them in the way they are currently represented, but at least a section that deals with the alternative - that is how to represent reactions without reaction elements (which should encourage the use of metadata that describes the participants and their roles, etc). In the OWL modelling context there is a formal way to depreciate modelling constructs, but I don't know of a similar way for general xml info sets. I'm not set on this subset specification being 1.1.1, but I couldn't offer another alternative. If it were code I would probably have gone for something like 1.1.1a1 - I think at least maintaining the 1.1 is important since the resulting models are 1.1 compliant. > > Andre. > _______________________________________________ > cellml-discussion mailing list > [email protected] > http://www.cellml.org/mailman/listinfo/cellml-discussion > _______________________________________________ cellml-discussion mailing list [email protected] http://www.cellml.org/mailman/listinfo/cellml-discussion
