Forums

> Parameters have no default units

No. But their unit is optional. Therefore, without explicit definition,
since the ratelaw gives substance/time and the species is substance/size,
we know that "nuc * cyt" is size/time.

> Look, my point is just that this has not (yet) demonstrated
> an ambiguity in the definition of SBML's identifier
> namespace interpretation.

And what do you think I'm doing here? Let-me phrase it in another way:
Marco and I cannot correctly render SBML rate-laws in BioModels. We don't
want to print "LDKJ9877 * 2767gbHGG * 9898ddd" but "kon * ATP *
cytoplasm".
In order to do that, we want to replace all the "ids" by the corresponding
"names" when they are present. And there is currently no way to
unambiguously do that, except by relying on the cleverness of the
SBML-export software, and the good behaviour of the user of the software.
It is to be noted that the user of a modelling software very rarely knows
the internal of the software, except if he developed it.

Let's make a poll right-now: To the people who develop SBML-export: who
checks that the id of a local parameter does not overide the id of a
compartment or a species? And for those who do, how do you solve the
problem?

> What it does point out is that
> software has to be designed more carefully than might first
> appear.

No. It points out that all the software in the world use different
namespaces for species, compartments and parameters, and we should have
though of that before.

> And I'm not against the idea of entertaining a change in
> identifier namespaces, if people think this is a problem
> that should be dealt with in SBML. We can put it to a
> vote if enough people say so.

I do. I think we should think more deeply about the problem. In
particular, I think we should challenge the various solutions, and push
them to the limits (In general I think we should do that for every
change/evolution of SBML)

Do anybody feel the same beside the one wo already said so?

--
Nicolas LE NOVÈRE, Computational Neurobiology,
EMBL-EBI, Wellcome-Trust Genome Campus, Hinxton, Cambridge, CB10 1SD, UK
Tel: +44(0)1223 494 521, Fax: +44(0)1223 494 468, Mob: +33(0)689218676
http://www.ebi.ac.uk/~lenov