Re: Links and graphs

KK:

thanks for the quick reply. i agree that time is very short for working
through the details of some of these problems.  thanks for taking it a step
further here. i recall reading through your presentation earlier this year,
so some of this rings bell in my head.

my primary observation here is that, in case one, it's a "guess" (e.g. "it
would probably have a VoID description"). this seems to be a very common
case. M2M clients will need to be prepared to make a "guess" and evaluate
the results. this seems to call upon the "gulf of execution" and "gulf of
evaluation" concepts of Norman[1] and his seven stages of action[2].

AFAICT, gaming systems use this pattern, but hypermedia and RDF systems
rarely do. any pointers to the contrary are most appreciated.

the second case seems to be an optimized version of the first, right? just
leaves out the first "guess"

Cheers.



[1]
http://www.interaction-design.org/encyclopedia/gulf_of_evaluation_and_gulf_of_execution.html
[2]
http://2.bp.blogspot.com/_y28DEEK8ElM/TBXcRAfW9RI/AAAAAAAAAEA/C74wNy6BrPo/s1600/UserInterfaceDesign0101.gif

mca
+1.859.757.1449
skype: mca.amundsen
http://amundsen.com/blog/
http://twitter.com/mamund
https://github.com/mamund
http://www.linkedin.com/in/mikeamundsen




On Sun, Dec 16, 2012 at 3:07 PM, Kjetil Kjernsmo <kjetil@kjernsmo.net>wrote:

> On Sunday 16. December 2012 14.53.09 mca wrote:
> > thanks for the jumping in here. my only complaint w/ your reply is that
> > you "described to me" how you would express this choice, but didn't show
> > me. i am serious, too.
>
> Fair point! I have one proposal (that needs elaboration, for sure), in the
> presentation I gave at the LAPIS workshop on ESWC:
> http://folk.uio.no/kjekje/2012/lapis2012.xhtml#%2831%29
>
> To do this in practice, it would first look for the pattern
> </resource> void:inDataset ?dataset.
> If it finds that this doesn't match anything, it simply cannot tell whether
> a SPARQL endpoint is present. If ?dataset is a URI, it should GET
> that/those resource(s), then it would probably have a VoID description.
>
> In the merged graph of the resource description and the VoID description,
> it would look for a pattern
> </resource> void:inDataset [ void:sparqlEndpoint ?endpoint ]
> if it finds that it matches, then there is a SPARQL endpoint.
>
> Obviously, I should sit down and write the code, but alas, there is too
> little time in the world. :-/ I've set myself a deadline of February to
> code the server side things for this though.
>
> Kjetil
>
>