Re: re-opening the DESCRIBE issue

On Mon, Apr 03, 2006 at 11:39:20AM -0400, Bijan Parsia wrote:
> On Apr 3, 2006, at 11:24 AM, Dan Connolly wrote:
> 
> >On Mon, 2006-04-03 at 11:11 -0400, Bijan Parsia wrote:
> >>On Apr 3, 2006, at 10:56 AM, Eric Prud'hommeaux wrote:
> >>[snip]
> >>>I propose the following text:
> >>>[[ http://www.w3.org/2001/sw/DataAccess/rq23/#describe
> >>>10.4 Descriptions of Resources
> >>>
> >>>Current conventions for DESCRIBE constrain it to return an RDF
> >>>graph.
> >>
> >>You mean the specification? Isn't there reason to leave that a bit
> >>open, e.g.,
> >>	DESCRIBE according to CBF as HTML
> >>
> >>I.e., why put in *this* constraint (convention?)?
> >
> >Part of the reason is that EricP is getting sorta conflicting
> >advice and trying to work a bit rapidly.
> >
> >If you have words that you'd prefer, I'd very much like to see them.
> >
> >If you don't, I can understand.
> 
> How about:
> 
> DESCRIBE is currently not constrained at all. The Working Group felt 
> that the abstract functionality was useful, but there was little field 
> experience to guide standardization. The most common understanding of 
> DESCRIBE is that it returns an subset of the queried graph that is 
> relevant to the resource, but the nature of the relevance function has 
> not been settled. Furthermore, it's possible that other relations 
> (e.g., metadata not in the graph, statistics about the resource, etc.) 
> and other return formats (HTML) may be part of the DESCRIBE form.
> 
> Hmm. This is all non-normative, except maybe the first sentence. 
> Frankly, I think that is as it should be. Since it really *isn't* 
> specified, even to the small degree hinted, we shouldn't pretend it is.
> 
> Cheers,
> Bijan.

Sorry to respond so late...

Side-band conversations with the director indicate he is (would be?
how vaguely must I speak of this?) happy with the current text in the
editor's draft:

[[
10.4 Descriptions of Resources (Non-normative)

Current conventions for DESCRIBE return an RDF graph without any
specified constraints. Future SPARQL specifications may further
constrain the results of DESCRIBE, rendering some currently valid
DESCRIBE responses invalid. As with any query, a service may refuse to
serve a DESCRIBE query.

...[original DESCRIBE text]...
]]

Some of the constraint text is from your suggestion. I phrased it as a
warning at the top of the section in order to still provide guidance
to folks who will use DESCRIBE anyways. In hind-sight, I wish I'd said
"specifi*c* constraints" as having the specification say something is
unspecified is a nearly vacuous truth.

Oh well, DAWG already ratified this text. I hope you find it acceptable.
-- 
-eric

office: +81.466.49.1170 W3C, Keio Research Institute at SFC,
                        Shonan Fujisawa Campus, Keio University,
                        5322 Endo, Fujisawa, Kanagawa 252-8520
                        JAPAN
        +1.617.258.5741 NE43-344, MIT, Cambridge, MA 02144 USA
cell:   +81.90.6533.3882

(eric@w3.org)
Feel free to forward this message to any list for any purpose other than
email address distribution.