[DAS] categorize

[Thomas Down]

FYI, Dazzle does respect categorize=(yes|no):

          view-source:
http://www.derkholm.net:8080/das/hsa_54_36p/features?segment=22:30000000,30010000;type=transcript;type=translation

          view-source:
http://www.derkholm.net:8080/das/hsa_54_36p/features?segment=22:30000000,30010000;type=transcript;type=translation;categorize=yes

...but it's always struck me as fairly useless -- one "optional" field in a
fairly verbose XML schema, and I certainly won't mourn it if you feel like
deprecating.

                   Thomas.

On Tue, Aug 31, 2010 at 5:15 PM, Andy Jenkinson <andy.jenkinson at ebi.ac.uk>wrote:

> Hi list,
>
> One of the (hopefully final) questions over the 1.6 specification concerns
> the "categorize" parameter in the features request. In DAS 1.53, this
> parameter is optional and may be either "yes" or "no". The default value is
> "no". The purpose is to instruct the server as to whether it must include in
> its response the optional "category" attribute of the <TYPE> element.
> However in practice, I believe most servers actually ignore this parameter
> and always include category information if it is available. If it is not
> available, servers cannot satisfactorily comply with a "categorize=yes"
> request and generally always provide an empty category attribute.
>
> In current drafts of 1.6, the category attribute remains optional but the
> categorize parameter has its default reversed (that is, it matches what
> servers currently do, and a request to specifically remove the category from
> the response can more easily be honoured by core server software. However,
> it is apparent that the functionality offered by the categorize parameter is
> really rather limited. Its only real purpose in this form is to reduce the
> volume of the response. In addition, given the introduction of the ontology
> and expansion of DAS to more diverse domains, the category itself is
> becoming less important. If, for example, the ontology cvIds become
> mandatory in the next version of the specification, we could feasibly
> deprecate it.
>
> So, it has been proposed that the categorize parameter be deprecated in
> 1.6, with a view to also deprecating the category attribute if and when
> ontology support becomes mandatory. Does anyone have any objections?
>
> Cheers,
> Andy
> _______________________________________________
> DAS mailing list
> DAS at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/das
>