Re: [GRAPH] graph deadlock?

On Dec 17, 2011, at 18:21 , Pat Hayes wrote:

> 
> On Dec 17, 2011, at 9:37 AM, Ivan Herman wrote:
> 
>> (I am still on sick leave and will stay for a few more weeks. But I am emerging from my torpor, and what better way of doing that is to read through a thread on named graphs?:-)
>> 
>> I have gone through the thread of the past few days, and I have this feeling of déjà vu. Although there are some neat things that came up (I like Sandro's GSR-s), I have the impression that we are in a deadlock.
>> 
>> Maybe we have to accept the deadlock and move on from there. Namely: I have the feeling that we have two notions here, and the group (and the community, probably) is just too divided to reconcile those. So, maybe, we have to accept that we have two distinct notions here; we should document both in the RDF concepts, and try to live our lives. The two notions that I see are
>> 
>> (1) RDF Datasets. It consists of labelled graphs: (G, l), where l is an URI. (Some raised the possibility to use literals for 'l', but I think there is a consensus to use URI-s). There is no semantic relationship between 'G' and 'l', so something like (with an ad-hoc syntax here):
>> 
>>  ( {a:b c:d e:f}, mailto:ivan@w3.org }
>> 
>> is a perfectly o.k. labelled graph in an RDF Dataset
>> 
>> It seems that most (all?) quad stores fall into this category as well as the datasets in SPARQL
>> 
>> (2) Named Graphs. It is a special RDF dataset, where the label 'l' is a (HTTP?) URI with an additional behaviour: if that URI is poked (GET-d) then it results in the serialization of a Graph whose parsing yields an equivalent graph to 'G'. It is the right/good framework for, say, Linked Data, etc.
>> 
>> Obviously, details are to be filled in.
>> 
>> My impression is that some in the group (Richard? SteveH?) would like to have only (1), whereas others (Sandro, Pat?, TimBL?) would like to have only (2). This has been the core of the discussion for 6 months. Let us face it, we will _not_ have a consensus here. 
> 
> Fine so far, but we get into a problem when we do the details. I personally have no problem (well, a kind of private horror, but thats just me) with datasets as you describe them here. The problem is, people apparently want to both have these AND use the URIs in them to refer to the associated graph in RDF triples (eg in 'metadata' graphs.)


Ok. That may be a third version; but another way is to say that, in the case of the Datasets, we do not specify the usage of that label URI. And we stop there. Again due to a complete lack of consensus that I just do not see how we could overcome.

What you say is not in contradiction with my proposal: specify RDF Datasets and Named Graphs as two distinct notions along the lines outlined above, and stop there...


> And that combination is simply illegal, according to the 2004 RDF specs. We have had a barrage of suggestions for how to wriggle around this, basically by redefining (or ignoring) the semantics so as to make some RDF no longer be RDF, but none of these work. That is why we are stuck. This situation cannot be resolved simply by letting it all hang out. We could simply declare that RDF has no semantics, and is simply to be used by programmers to mess around with in ways they find handy. Really, this might be the best way to move forward. But until we do this, we have to take the semantics seriously. 
> 
>> 
>> Is it so bad if we acknowledge this, and we clearly document the two notions? Both notions can be documented in RDF Concepts, maybe some (although I am not sure) in the RDF Semantics. We may think about some extra RDF relations among probably named graphs (sub-graph, things like that) which probably do not make too much sense for labelled graphs. In any case, as long as things are clearly defined, implementations, applications, etc, may decide what they do, what they expect, etc, and that is fine for interoperability.
> 
> Actually, no it isn't. Would the Web work if HTML was specified this way? 

I am not sure I understand. An application may declare whether it uses Datasets or RDF Graphs. If Datasets, then only labels are used for URI-s, which I believe is the way quadstores, for examples, are used already, as well as SPARQL datasets. For Named Graphs that may be used, for example, in the case provenance, applications know about the extra behaviour of the label (if GET-t, then bla bla bla...) which is then required by the standard.

Ivan




> 
> Pat
> 
>> 
>> I am not sure how this affects serialization.
>> 
>> Cheers
>> 
>> Ivan 
>> 
>> 
>> ----
>> Ivan Herman, W3C Semantic Web Activity Lead
>> Home: http://www.w3.org/People/Ivan/
>> mobile: +31-641044153
>> FOAF: http://www.ivan-herman.net/foaf.rdf
>> 
>> 
>> 
>> 
>> 
>> 
>> 
> 
> ------------------------------------------------------------
> 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
> 
> 
> 
> 
> 
> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
FOAF: http://www.ivan-herman.net/foaf.rdf

Received on Sunday, 18 December 2011 11:35:45 UTC