Re: The way we do things in the Semantic Web community (was: Re: Labelled graphs)

See comment/question inline below.

On Apr 25, 2012, at 3:16 PM, Richard Cyganiak wrote:

> On 25 Apr 2012, at 19:44, Sandro Hawke wrote:
>>> Sandro, if you want to design brand-new stuff and try if it works, then you should change jobs and come work in academia for a while. We get paid to do research!
>> (That's a little close to ad hominem, Richard.) 
> (It's not meant as a personal attack; I wouldn't be saying this if you were representing MIT or an implementer or a group of users.)
>> But it's well known that once something enters the W3C process, no
>> matter how "battle-tested" it might be, the WG still finds lots to
>> improve.   So there's always some additional research/design/development
>> work that has to happen during the WG process.  
> Yes, some additional R&D is usually necessary.
>> I will grant that sometimes it turns out there's too much additional R&D
>> to do, and we need to re-think the WG, backing out of the W3C process.
>> This may be one of those situations, although we had enough consensus
>> today that I think we can go ahead with some of it, at least.
> I agree.
>> On the other hand, please consider my point: sometimes we can't know
>> whether a design will work until trying it in a fairly big arena, with a
>> lot of attention.  
> I question whether standardization in W3C is the right way of generating this attention. I believe it is not, due to the danger of poisoning the big arena with a poor design.
>> As I understand the history, XML was designed by a
>> W3C Working Group.  
> As I understand the history, XML was made by throwing away most of SGML. Subsetting an existing standard based on deployment experience is less risky than creating something new.
>> Do you think something has been blocking the research since 2005 but
>> would allow it to really move ahead now?
> I'm glad you asked. Let me quote what I wrote two years ago for the workshop that kicked off this WG's work:
> [[
> Managing context, provenance and graph updates are extremely important in almost any RDF application. The solution is the Named Graphs data model. It is already part of the SPARQL Recommendation, is widely implemented also outside of SPARQL, and generally well-understood.

So, in your view, the SPARQL specs define a well-understood and widely implemented model for handling context, provenance and updating? But surely if we have learned anything in this WG, it is that this SPARQL-defined machinery is not in fact used consistenly, has no clear semantics (that is, one which supports all the intended or claimed use cases), apparently mis-cites and mis-uses earlier work (it refers., indirectly, to the "named graphs" paper and uses its terminology, but apparently its users did not read that far, and so current practice sharply diverges from the ideas in that paper, as you have pointed out) and is not even well-defined as stated. So our job, as I see it, is to try to clean up this mess  and provide a clean model, or models, to rationalise the uses of this construct. I see Sandro's suggestion as part of this process. 

>  It should be elevated to a separate Recommendation. Besides codifying existing practice, this would be a welcome support for those practitioners who are trying to improve the general state of provenance tracking and metadata

Ah, I forgot metadata. Add that to the list. 

> on the RDF-based Web and who are currently fighting a somewhat uphill battle because of Named Graphs' relative obscurity. Furthermore, a Named Graphs standard could galvanize research on the upper layers of the Semantic Web stack, where the availability of rich context information, along with a standard model for its representation, is a key requirement.
> ]]
> Codifying existing practice, and putting W3C's clout behind something that we already know works well, is sometimes all that the community needs to proceed, IMO.

And what do we do when "existing practices" differ widely and are apparently inconsistent with one another? And when they, or the claims made for them (handling provenance, metadata and updating) go way beyond the conceptual model underlying our existing 2004 standarda and at times are incompatible with them?

Obviously, a rhetorical question. You see my point, I am sure. And I can guess your answer :-)

> There's more that could be said in response to your question, but I'll leave it at that for today.
> All the best,
> Richard

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

Received on Saturday, 28 April 2012 15:44:35 UTC