Re: [Miase-discuss] Example file & thoughts of implementation - Source - XPATH - time

[Nicolas Le N. <le...@eb...>]

Frank Bergmann wrote:

> The next issue I had with the dataGenerator and the accompanying mathML
> element. Lets recap, a data generator consists of:
> 
>  a list of variables(with id, taskReference, XPath)
>  a list of parameters(with id, name, value)
>  mathml
> 
> here the mathml should only include expressions of constants, parameters and
> variables. Could someone please elaborate again, why we need a list of
> parameters here? We cannot really reference them from somewhere else, and
> setting them up with id,name(and why a name in the first place?) and value,
> rather than using the value directly in the mathml seems strange to me. 

Having them outside and named not only increases the readibility, but also permit to put annotation. So we can for instance normalise a result using a value found in the literature, or at bionumbers.org and specify the provenance.


> Which seems incorrect to me, but even if reformulated in a more appropriate
> form:
> 
> <math xmlns="http://www.w3.org/1998/Math/MathML">
> 	<csymbol encoding="text"
> definitionURL="http://www.sbml.org/sbml/symbols/time">time</csymbol>
> </math>
> 
> It is still troublesome, as I'm not sure which "time" from which
> taskReference to take. So instead we should have time refered to in the
> variable section: 

Absolutely.

> <variable id="time" taskReference="task1" XPath="/sbml:sbml/sbml:model/time
> "/>
> Except this is not really a XPath element, and so should be replaced with
> ... what?

That's a very good point, but I think the time is not related to the model, but to the simulation definition. One could introduce a built-in symbol time in SEDML/SimML, and map it to simulation definition somehow.

If we want to use the SBML time, one way around would be to modify the models meant to spit out the time by adding a parameter "t" and and assignmentRule using the csymbol. But of course, that's a changeXML and we decided to postpone those. 

But really, I think the time is defined in the element uniformTimeCourse, not in the model.

-- 
Nicolas LE NOVERE, Computational Neurobiology, EMBL-EBI, Wellcome-Trust 
Genome Campus, Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521
Fax:468, Skype:n.lenovere, AIM:nlenovere, MSN:nle...@ho...
http://www.ebi.ac.uk/~lenov/, http://www.ebi.ac.uk/compneur/