Re: [Fwd: Comments on SPARQL] (entailment, soundness, completeness)

On Sep 20, 2005, at 12:12 AM, Enrico Franconi wrote:
> [...]
>> A SPARQL engine is not required to perform any inferences, to check 
>> semantic redundancy, etc. (Though it may, it is not required to.)
> This should be decided by the user, I guess.

Perhaps. Regardless, it's a very good point that is not getting a lot 
of bandwidth yet.
This WG should be considering which capabilities we're going to give to
users as much as the technical details of how to design those 

I think Pat's message was in some ways factually accurate about how we 
to the SPARQL last call design, but what I'm seeing comments that say 
we didn't take
all the right requirements into account when we did that design.

So I opened a couple issues, and all this technical stuff is in order, 
but I'd sure like
to see an equal amount about real-world use cases and such. The main 
I have seen so far are

  * building a SPARQL service out of RDF software components that lean 
their graphs
   (cf PFPS's comments)
  * using SPARQL to access graphs at the abstract syntax level in order 
to manipulate them
  (cf recent mail from AndyS)

I think Kendall summarized a similar tension well in a recent 
teleconference, but I'm not sure
it got minuted.

Enrico, elsewhere in your message about "Adoption of entailment in 
of September 19, 2005 11:55:09 PM GMT+01:00, you wrote "here we don't
argue whether this is useful and how this is going to be used." Note 
that I
pretty much stopped reading at that point. I'll be more motivated to 
study the technical
details when I know which user requirements, use cases, and applications
a proposal is intended to address and which it's not intended to 

Other people might prefer to read the technical details first, and 
email is
happily asynchronous. So don't let me stop you writing about whatever 
feel like writing about first. But just keep in mind that before I 
think we're
ready to close these issues, I'll need to see more discussion of user

Dan Connolly, W3C

Received on Tuesday, 20 September 2005 11:04:01 UTC