W3C home > Mailing lists > Public > public-rdf-wg@w3.org > April 2012

Re: some GRAPHS strawpolls for today (agenda?)

From: Pat Hayes <phayes@ihmc.us>
Date: Sat, 28 Apr 2012 16:16:56 -0500
Cc: sandro@w3.org, public-rdf-wg@w3.org
Message-Id: <DA2F5BDB-D7BB-4B0F-9C6D-7BE7379768BE@ihmc.us>
To: William Waites <wwaites@tardis.ed.ac.uk>

On Apr 28, 2012, at 11:09 AM, William Waites wrote:

> On Sat, 28 Apr 2012 10:57:00 -0500, Pat Hayes <phayes@ihmc.us> said:
>    phayes> As I see it, there are two "models" for this whole issue
>    phayes> on the table. In yours, a dataset is effectively a
>    phayes> graph-baptism device for attaching names to graphs. In the
>    phayes> other, the term "graph name" is a misnomer, and the fourth
>    phayes> field is more like a context or 'extension' name, in any
>    phayes> case one which can influence the interpretations of the
>    phayes> IRIs in the graph. Call these respectively the Name and
>    phayes> Context views. Then different views give different votes.
> I've pointed out before that I can see people wanting to do *both* of
> those things.

I am sympathetic to this position. But in this case, we need to have TWO DISTINCT mechanisms or syntactic devices to distinguish the two cases, because they are about as distinct semantically as two things can get. If we go on phrasing this discussion as "what is the right way to interpret Trig syntax", then we will never reach a stable consensus.

> So when we look at proposals about graphs and compare
> them to the use cases we have on the wiki, we look and say whether or
> not a proposal can work with each of the use cases. But do we look to
> see if a proposal can simultaneously work for more than one or even
> (ideally) all the use cases?
> For my part, I'd like a way to talk about context and provenance and
> such *and* be able to partition or group statements in arbitrary ways
> that might, for example, be useful for putting them on web pages. If
> these are incompatible uses of the fourth column than we're in
> trouble.

Then we are in trouble (hardly news :-) since yes, these ARE incompatible. (I cant help remarking that it is taking a very long time for people to grok this rather basic point. Isn't it blindingly obvious? If you have a piece of syntax which you want to (a) mean something sharp and definite and also (b) be freely used to mean something completely different, then you have put yourself up the creek without a paddle. )

BTW, part of the thinking behind the 'contexts' or 'extensions' idea I recently tried to outline was to allow Sandro's 'naming' semantics to play the role it does, and add the 'contexts' idea to it by a different mechanism, trying to make sure they play reasonably well together.

> I'll just briefly mention that if we use the fourth column to store
> statement identifiers we stand a chance of building up all of the
> machinery from there, but that suggestion seems to get consistently
> shouted down so I won't push it...

I dont even know what that means, but it sounds like another way to say Sandro's idea. 


> Cheers,
> -w
> --
> 	    William Waites MBCS <wwaites@tardis.ed.ac.uk>
> Visiting Researcher, Laboratory for Foundations of Computer Science
> 	    School of Informatics, University of Edinburgh

IHMC                                     (850)434 8903 or (650)494 3973   
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
Received on Saturday, 28 April 2012 21:17:31 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:04:16 UTC