Forums

>> Maybe I'm wrong in this, but it seems to me we want a
>> way to distinguish the intention behind the annotation.

lenov> Yes, yes, you're right. Although I wanted to defend
lenov> our scheme, I vote for a CV attribute in the
lenov> present case. I don't think SBO terms are really
lenov> annotations. They will be true part of the SBML
lenov> files.

After some more thought and (failed) experimentation with
alternatives, I'm beginning to come around to this idea.

In fact, what about adapting the (similar) notion from
MathML, and calling it 'definitionURI' ? In MathML, most
items are permitted to have a 'definitionURL' attribute
(which is actually a URI -- it's misnomer and they say so).
The meaning of the definitionURL is:

4.3.2.3 definitionURL

csymbol, declare, semantics, any operator element

Points to an external definition of the semantics of the
symbol or construct being declared. The value is a URL or
URI that should point to some kind of definition. This
definition overrides the MathML default semantics.

In SBML, the analogy is that a definitionURL would allow the
model to point to an external clarifying definition of the
thing in question. This would work well for reaction rate
equations, and it might actually work well for other things
like species. It would mean that the SBML core structure is
in some sense the common denominator, lacking a lot of
detail but interoperable between tools, and the
definitionURL on elements could point to additional details
about the thing in question. It would be more than an
annotation (which is in some sense informational and
optional). It would allow a tool to clarify the intention
behind a given object ("I mean this reaction 'foo' to be
this specific type of reaction here", "I mean this species
'bar' to be the specific chemical compound known as 'ATP'",
etc.)

Does anyone else have comments on this?

This is not a formal proposal, but I'll write it up if no
one thinks it's egregiously flawed.

MH