Hi all,

The subject of units and associated issues has been popping up quite a 
bit lately, and a lot of the current problems are mainly due to 
publication of models without full dimensional annotation and/or errors 
in the initial encoding of a model into CellML. From my experiences, I 
have found that when a model is accurately and completely units balanced 
from inception to completion, then a lot of the issues with units are 
not apparent.

In order to help overcome the initial hurdle in assigning units to 
variables when the reference source my not be complete in regard to 
units, Erik Butterworth of JSim fame has proposed quite a neat solution. 
Currently, as part of the units checking in JSim, JSim will try to 
assign units to variables whose units are not assigned. This process is
reasonably accurate assuming there is a base of variables whose units 
are correctly assigned. In this example (in MML syntax):

real A mm;
real B sec;
real C;
real D = A/(B+C);

JSim would assign C units of sec (to match B), and assign D units
of mm/sec (to match A/(B+C)).

This feature can currently be used by coding up models in MML first and 
then grabbing the generated units for use in the CellML encoding - a 
fairly awkward process. What Erik is proposing is:

"One possible project would be for JSim to read in CellML files which 
might be missing unit assignment, and add them in as necessary.  JSim 
could then either proceed with modeling, or write out the unit enhanced 
CellML and exit.  This approach would probably take a few weeks on this 
end (assuming I can budget the time), but I could consider it if your 
team thinks they'd like to work with such a tool and integrate into the 
CellML archival process."

I am trying to determine if there would be enough interest from CellML 
model authors to make it worthwhile pursuing such a project? Anyone 
interested please speak up :-)


David.

-- 
David Nickerson, PhD
Research Fellow
Division of Bioengineering
Faculty of Engineering
National University of Singapore
Email: [EMAIL PROTECTED]
_______________________________________________
cellml-discussion mailing list
[email protected]
http://www.cellml.org/mailman/listinfo/cellml-discussion

Reply via email to