Finalizing L1v4: Dependent Variables

[Lucian Smith]

According to:

https://docs.google.com/spreadsheets/d/1-7uRLppjiGZ1jyhEDRhpWLDi9iR9Oi9cBQjplv0OBnk/edit?ts=605a2d23#gid=0

there are only a few unimplemented things in L1v4 left, that we can either just ditch, or try to drum up implementation for.

As far as Tellurium and I go, the one thing left that I plan to do that I haven't done is parse the expanded DependentVariable syntax.  However, I need more information if I am to do so.  The notes say that they're used for "accessing control coefficients" and "time-dependent sensitivities".  What does this mean, specifically?  Is there a list of the symbols involved so that I can implement support for them?  Do I need to come up with a list myself?

-Lucian

[Frank Bergmann]

Hello Lucian, 

the idea was, that the list for the DependentVariable would be on the Specifications website, so it could be readily enhanced by pull requests, when more than one tool would support a certain additional term. I would expect the list to be of the following format: 

      term | general description | meaning of symbol | meaning of target | meaning of symbol2 | meaning of target 2 

Since we introduced the construct, we also changed the meaning of Variable to allow both `symbol` and `target`, so there is some overlap there. I still think we should also have a list for allowable symbols there. (For example the rate of change of a species as in the spec example could now be implemented using solely a variable with both symbol, and species target. DependentVariable would be needed for the case, that an additional target would need to be specified. 

Another question is how the terms would actually look, back when we introduced it, i was assuming them to be of the form: 

      urn:sedml:symbol:

though it would seem that others prefer KISAO terms for it. No matter what form they will take, we will need the description of what the symbols / targets mean. 

I hope this helps, 

cheers

Frank

[Lucian Smith]

This is great!  It matches the general idea of what I was thinking myself.

However, it looks like nobody has actually come up with this table yet, or even what terms should be on the table?  I've seen a few proposals over the years, but I guess they were never collected in one place.  So, I created a place for them:

https://docs.google.com/spreadsheets/d/1f3rBUUhOsB6k7-i_ins0nlB7PTtd_huvhgFrCOh501U/edit?usp=sharing

I put things on it that I think would make reasonable additions, but this should be considered a starting point, not anything final.

One thing we could add would be the ability to get individual elements out of some of the matrices--say, 'the Jacobian value for reaction J1 and species S2', with those two elements set as target and target2.

Given our newfound influence over KiSAO terms, I am inclined to suggest that we use KiSAO for most if not all of these symbols.

-Lucian

--
You received this message because you are subscribed to the Google Groups "sed-ml-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sed-ml-discus...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/sed-ml-discuss/a81a2020-18b9-4b64-a281-9e08346d84d6n%40googlegroups.com.

[Jonathan Karr]

Happy to help with the KiSAO part, or orient anyone who'd like to volunteer to do it. Editing KiSAO is straightforward once we've decided how to organize it. 

Actually, I think it's unnecessary to worry too much about the organization of terms. At least for now, the key thing is unique terms and descriptions of each one. More relationships among the new terms could be added later with little to no impact on the initial usage.

Jonathan

To view this discussion on the web, visit https://groups.google.com/d/msgid/sed-ml-discuss/CAHLmBr2ESwptXhhCk5VAv4NOmr0pYHytQ0ZoLcNpBDC4gLNZeQ%40mail.gmail.com.