Re: [Psidev-ms-dev] PSI-MSSWG call in 9 hr

[Angel P. <an...@ma...>]

On Tue, Mar 25, 2008 at 2:49 AM, Eric Deutsch <ede...@sy...>
wrote:

>
> - Address cvParam category name issue, as described in spec doc. Do we
> want:
>
> A) <cvParam cvRef="MS" accession="MS:1000583" name="SRM spectrum"/>
>
> C) <cvParam cvRef="MS" accession="MS:1000583" name="SRM spectrum"
> categoryAccession="MS:1000035"/>
>
>  (where MS:1000035="spectrum type")
>
> The category accession is not needed or used under normal circumstances by
> Darren's reader and presumably most readers. But it could be useful in cases
> where a term not known to the reader is used. Under scenario A, if the
> reader software predates MS:1000583, the reader will be very hard pressed to
> know what to do with this term. In theory, under the same scenario with
> option C, the reader could more easily determine what to do with this piece
> of information even though the exact term is unknown. Maybe "cannot
> determine spectrum type" should be a terminal error, while "unrecognized
> spectrum type" might merely be a warning. Darren and I discussed for a while
> at US HUPO. Let's see if we can come to a decision.
>
> Can't attend the call, but I really can't keep quiet about this.
categoryAccession, or any other method of scoping terms (or essentially
defining new terms) within xml instances that do not live within the
referenced CV itslef is a *bad* idea. Here's the thing about cvRef, you can
use more than one source wihtin an XML instance. Here is the thing about
CV's and ontologies, they can reference and inherit from each other.

So option (D) is born:


D) <cvParam cvRef="MyMS" accession="ME:0000501" name="SRM spectrum"/>

and within the MyMS CV, we would have  ME:000501 IS_A MS:1000035

And the world is at peace and one with itself. Or at least that is the way
is should be.