[cellml-discussion] Should groups be allowed to imply metadatainformation?

[David Nickerson]

Andrew Miller wrote:
> Hi all,
> 
> The CellML 1.1 specification says:
> 
> "
> 6.5.3  Groups must not imply metadata information
> 
> Modellers must not use CellML groups to associate properties or 
> classification information with sets of components. The metadata 
> functionality is the proper method for making such associations. This 
> increases the chance of that information being used by a range of CellML 
> processing software.
> "
> 
> If extension groups cannot be used to imply metadata or mathematical 
> information, then there is not really anything left for them to imply. I 
> think that we should do one of the following:
> 1) Non-standard relationship types be disallowed, and only encapsulation 
> and containment be kept (encapsulation does affect the mathematical 
> formulation of the model, while containment is really metadata 
> information), or perhaps only encapsulation should be kept, with 
> containment data represented in metadata, or,
> 2) Allow groups to be used for metadata information, but in the 
> informatively annotated specification encourage the CellML community to 
> standardise on exactly how a certain type of metadata should be 
> represented (this is required whether RDF/XML or groups is used to 
> express the metadata anyway).

I'd be in favour of option 1 combined with moving containment into metadata.

Just to add a link to the tracker, also see: 
https://tracker.physiomeproject.org/show_bug.cgi?id=316