David Nickerson wrote:
>> I'm not sure we would name them so the URL needed a query portion, 
>> because that would mean that browsers would save the model wrong. A 
>> hierarchy (perhaps a subset of an existing ontology restricted so it 
>> becomes a tree) could be used to locate the model from the path rather 
>> than query, e.g. 
>> /models/channels/sodium/ventricular/human/tenTusscher.xml might be 
>>     
>
> so you are suggesting that we move away from a flat directory structure 
> or is your example above addressing an ontology which would actually 
> resolve to a URL like 
> http://www.cellml.org/models/2004_ten_tusscher_noble_noble_panfilov_version02_part07.xml?
>   

I think that having only one URL for each model is a good idea, because 
that makes it much easier to define metadata about a model outside of 
the model (otherwise, there are two equivalent URLs, but software 
doesn't know they are equivalent unless it does some extra work to do 
this), so if our URLs are not informative enough, perhaps URLs with 
ontological information in them should be used for library components.

The current naming system works well for flat models (all models are 
currently flat, because we don't support CellML 1.1 in the repository), 
and so I think that it should be preserved for top-level, non-reusable 
models (these would presumably correspond to your experiment CellML 
files). Keeping top-level models with the current naming system also 
ensures that we don't rename any existing models.

Library models have never really been addressed by the naming 
specification, and I think that some sort of hierarchical naming system 
like my example above (but probably with year information as well as the 
author on the final component of the path) would be a good approach. The 
only difficulty is deciding which ontological terms to include in the 
names, as this likely varies depending on the highest level terms.

Best regards,
Andrew

_______________________________________________
cellml-discussion mailing list
[email protected]
http://www.cellml.org/mailman/listinfo/cellml-discussion

Reply via email to