Re: SPARQL EXISTS

good morning;

i had corresponded with adrian earlier this week about the exist issue and he had asked me to add a comment to #156 as a reference for meeting participants.
i declined to do that because, on principle, that issue should not exist.

i responded, instead, directly to adrian and ora with a statement.
i elaborate on it here, below, with a text which i ask to be included in the meeting minutes by reference or to be permitted the opportunity to read it into the minutes during the meeting.

---

i intend that during this meeting the group address the concern, that any decision to change the sparql recommendation to specify an interpretation of the exists operator which goes beyond the text already present in the current recommendation is beyond the scope of the working group charter.
i do not intend to discuss the merits of proposed operator implementations.
this despite that the matter is important and that the language definition should change to address it.
the problem is, the rdf star working group is not the correct venue.

this in both practical and formal terms.
on one hand the rdf star effort has already been going for close to five years in various forms without yet achieving an adequate approach to annotating statements.
on the other hand, in formal terms, while the charter does permit to address "pending editorial errata", the suggested changes are well beyond that.

even if one might consider that they constitute a "new feature ... in order to render future evolutions easier", that requires a very generous comprehension of that provision and such changes are to follow a process which requires public review and which involves following a technical report from draft to recommendation.
no such process is apparent in this case.
there was a community group effort towards a change to the exist operator definition several years ago and that group did (to my recollection) formulate the approaches which have reappeared as the current proposals.
that group's report was not conclusive and the editor confirmed that by removing the indication that a consensus had been reached in the final report draft.

even the comments on #156 confirm that this matter is not an editorial errata.

the underlying issue is that the upwards scoped lexical bindings which are central to the language definition are not only not adequate to implement exists, but also not sufficient to support other essential features. among them, as called out in the #156 comments, lateral joins and query parameters.
changes of this sort are certainly not editorial, but rather on the order of additional features, which the charter would allow, but for which it indicates a different process than the resolution of a github issue.

this does not concern the merits of specific approaches to the operator definition, neither those of the two approaches identified in the agenda, nor of alternatives.
this working group is not in a position to address the broader problem and should not presume a solution which limits the future broader response.
that work is not among this working group's tasks.
the most it can do is propose a charter for a group which will have participation and time adequate to work on the topic.

---

best regards, from berlin

> On 3. Oct 2024, at 10:33, Adrian Gschwend <adrian.gschwend@zazuko.com> wrote:
> 
> Hi everyone,
> 
> FYI there were some comments this week on the issue. As we have it in the agenda tonight, I encourage everyone to take some time to check out the answers by Pavel/Stardog and Hannah & Johannes/QLever in preparation for our discussion:
> 
> https://github.com/w3c/sparql-query/issues/156
> 
> regards
> 
> Adrian
> 
> -- 
> Adrian Gschwend
> CEO Zazuko GmbH, Biel, Switzerland
> 
> Phone +41 32 510 60 31
> Email adrian.gschwend@zazuko.com
> 
> 

---
james anderson | james@dydra.com | https://dydra.com

Received on Thursday, 3 October 2024 10:32:58 UTC