describing the algorithm used to compute the DESCRIBE result set

Hi all,


better late then never I'd like to propose the following additions to DESCRIBE :

1) to add an identifier INFORMAT to the resulting description which identifies the algorithm used to construct the description.

2) to add a modifier ASFORMAT to the DESCRIBE operator, which is used to describe the algorithm that the querying service would prefer.

3) The use of both INFORMAT and ASFORMAT should be optional. The algorithm must be identified by a (cool) IRI.


because: 

(1) would help to automatically make sense of the data returned by a service and automagically reformat or reuse it in appropriate ways.
It would also facilitate and promote the reuse of ontology patterns, the support of common usecases, the establisment of modelling primitives and the explicit support for special vocabularies (like the blank node heavy FOAF).

(2) is to the server side what (1) is to the client.

(3) There's no need to make INFORMAT and ASFORMAT mandatory. In absence of a INFORMAT identifier querying services will have to do what they already do: somehow try to make sense of the returned data. In absence of a ASFORMAT modifier the Sparql endpoint can return its preferred/default format just like it does now. 


Future additions may provide more elaborate constructs by which a service can advertise the formats it offers and by which a client can specify which formats it prefers and which formats would be okay as fallbacks.


Thanks for considering this and sorry if a similar propsal has already been discussed!
Thomas Lörtsch

Received on Monday, 19 July 2010 18:21:51 UTC