W3C home > Mailing lists > Public > public-rdf-wg@w3.org > October 2011

Re: proposal to close ISSUE-77 (Re: [ALL} agenda telecon Oct 19)

From: Andy Seaborne <andy.seaborne@epimorphics.com>
Date: Wed, 19 Oct 2011 13:32:15 +0100
Message-ID: <4E9EC34F.4070606@epimorphics.com>
To: Sandro Hawke <sandro@w3.org>
CC: public-rdf-wg@w3.org


On 19/10/11 13:17, Sandro Hawke wrote:
> On Wed, 2011-10-19 at 11:23 +0100, Andy Seaborne wrote:
>>
>> On 19/10/11 01:44, Sandro Hawke wrote:
>>> On Tue, 2011-10-18 at 21:05 +0200, Guus Schreiber wrote:
>>>> Maybe we should have resolved at the FTF to skip a week after all the
>>>> hard work at the FTF, but we didn't, so here is the agenda:
>>>>
>>>>      http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2011.10.19
>>>>
>>>> I suggest a short meeting, basically reviewing the FTF minutes, with
>>>> some reflections, plsy tackling two detailed issues.
>>>
>>> The second of those is ISSUE-77.   To sum up the discussion on the
>>> mailing list, I think our best chance for consensus is on this
>>> resolution:
>>>
>>> PROPOSED: Close ISSUE-77 with a plan to keep rdf:Seq and RDF Collections
>>> as in 2004 (syntax, no semantics), but include non-normative text in one
>>> or more of our documents gently steering people toward best practices,
>>> which are (1) try to model without using either one, when feasible, and
>>> (2) if you need to use one, use RDF Collections structured so they can
>>> be serialized losslessly in Turtle using the "(...)" notation.
>>>
>>> This proposal is trying to split the difference: some people (including
>>> me) think it would be better to say something stronger (I'd like to
>>> deprecate Seq, or at least label it merely a "compatibility feature");
>>> some people want something weaker (like do nothing).  My sense from what
>>> people have posted is that everyone can live with this middle ground.
>>>
>>>      -- Sandro
>>
>> ISSUE-77 is specific to Seq's and I'd like to deal with Seqs and lists
>> separately because I feel we have to leave something in the docs that
>> can be used for ordering and isn't spoken against in some "don't use if
>> you can" way.
>
> I can't quite follow what you're saying.   I combined Seq and lists
> because people said "we can't get rid of Seq unless we tell people what
> else to use".   You say you want to deal with them separately, but ...
> then it seems like you're saying we have to leave in one or the other --
> which sounds like we have to deal with them together.

The text proposed says:

 >>> (1) try to model without using either one, when feasible

which I took as also pushing RDF lists into the "not preferred 
category".  That's seems to be letting the technology influence the 
modelling too much.

The fact that an ordering construct is an RDF list which "has issues" 
isn't the fault of the modeller.

>> I don't mind how what we do to rdf:Seq but if we say "use blank nodes
>> for Seq" (which then avoids the merge issues) it is a step forward (Ian
>> -- skolemized system generated URIs would count as well)
>
> I can live with that, but I'm not sure why we'd say
> dont-use-non-blank-nodes-for-Seq any stronger than dont-use-Seq.

It avoids merge problems as the bNodes should stop two rdf:_1's on the 
same resource.

>> Having gone back to the text around RDF Collections, some tidying up and
>> bringing together would be helpful although the primer is in reasonable
>> shape already.
>
> Oh, yeah, I'd also *love* to stop using the terms "collections" and
> "containers" and just use "Seq" and "List".   I know "Collection" is
> hard-coded into RDF/XML, but still.    Expecting people to remember
> container=Seq and collection=List is a bit ... rude.    Especially if
> we're calling g-boxes "Graph Containers", which I've already seen one of
> us mis-write as "RDF Container".

+1 - Absolutely agree

	Andy
Received on Wednesday, 19 October 2011 12:32:58 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:46 GMT