[DAS] next "feature" for 1.6 spec

[David Sherman]

I'd like to de-lurk here to say that this would be a great addition to
the spec. Currently we get this functionality on the client side by
sending a series of (geometrically) widening requests to get a feature
of the desired type on either side. Even with aggressive caching there
is a price to pay.

On Jun 29, 2009, at 15:44, Lincoln Stein wrote:
> There is an ambiguity here. It is completely possible for a DAS  
> server to
> return multiple distinct feature types per segment. However, a type
> "NEXTLEFT" does not distinguish which feature type can be found to  
> the left.


This would be a problem. We use next-left and next-right for navigation
on summary pages for individual elements, and the notion of "next"  
*does*
depend on the type of element (individual gene, locus, repeat element,
etc.). Reusing the type to indicate that this feature is provided even
though it is outside of the requested segment, seems abusive to me.

I suppose you could extend the feature ontology to include a new kind of
region to wrap around the next flanking features of each kind, and a
supporting server would add them on the fly. (What would you call this,
"speculative neighborhood"?)

Cheers,

djs               David J. Sherman          (David.Sherman at INRIA.FR)
                   INRIA équipe Magnome,  CR INRIA Bordeaux Sud-Ouest
                   Laboratoire Bordelais de Recherche en Informatique
                   voix : +33 5 40 00 6922     fax : +33 5 40 00 6669
                   icbm : 44°48'28.2"/-00°35'47.4"/49m