SPARQL Maintenance (EXISTS) Community Group [Was: Enough already]

There is a community group looking at the EXISTS issue.  Anyone can join.

The group have identified 5 technical issues and has two proposals under 
discussion.  There are new test cases as well.

One of the goals of the CG is to find a better definition while having 
an emphasis on maintaining compatibility with the places where the 
EXISTS definitions spec works.

It is only this one function, not the whole of the query language.

     Andy

To join:
https://www.w3.org/community/sparql-exists/

https://w3c.github.io/sparql-exists/

On 11/12/16 01:12, Terry Roach wrote:
> Martynas
>
> You are right, if there are fundamental issues with the design or approach (as you seem to feel), then of course they should be vigorously debated. But I can’t see that that is what is happening here. The discussions are quite argumentative and about seemingly trivial topics (at least to my untrained eye). These things should not be holding everything up.
>
> Peter
>
> Actually I think the SPARQL example is a great (and beautiful) one. We get enormous utility out of SPARQL, indeed our products and possibly our company couldn’t EXIST without it.
>
> A standard is only as good as its adoption. Where would we be if SPARQL was still sitting in the lap of some committee, arguing over the semantics of EXISTS?
>
> There is never any end to word smithing. Every author  would love to rewrite his work just as soon as the ink is dry.  Isn’t completely precise as illusory as completely perfect? What value does the last 10% of precision have and at what cost?
>
> Coming from a product mentality, to me the thing needs to get published with all it’s warts and wrinkles and get used. It can always be improved later and one day may become beautiful. Let the community be the judge and provide validation and feedback, while you keep on tidying up the precision for version 2.
>
>
> Terry Roach
>
>
>
>
>> On 11 Dec 2016, at 5:24 AM, Peter F. Patel-Schneider <pfpschneider@gmail.com> wrote:
>>
>> On 12/10/2016 03:52 AM, Martynas Jusevičius wrote:
>>> Hey Terry,
>>>
>>> as I see it, the debate is not about perfection, but precision.
>>
>> Indeed, overall perfection can never be a goal of a standard.  However,
>> standards do need to be precise.  As well, the standard does need to match
>> some notion of if not beauty then at least non-ugliness.  There also should be
>> some hope for utility as well.
>>
>>> Do you see people arguing about RDF or SPARQL specs? No, because they are
>>> defined precisely using semantics and algebra. That is the guarantee for
>>> robustness, not merely application in practice.
>>
>> SPARQL is not a good example here, as the specification of EXISTS in SPARQL
>> has multiple problems.  No SPARQL system implements EXISTS as specified.
>> Worse, different SPARQL systems implement EXISTS differently.  Of course, the
>> formal specification of SPARQL means that it is possible to determine just
>> when implementations differ from the specification.
>>
>> peter
>>
>>
>>
>>> LDP started a trend which SHACL seems to be following, that the editors are
>>> not able or willing to produce such precise, theory-backed definitions and are
>>> trying to push the spec out of the door ASAP. This might be of advantage to
>>> some players in the short term, but detrimental to the future of Semantic Web.
>>>
>>> In case of SHACL specifically, I think the problem is that while SPIN was an
>>> elegant concept on top of SPARQL, shoehorning constraints into a vocabulary is
>>> a model mismatch, a little like putting an ORM on top of RDBMS: it works most
>>> of the time, but there will always be corner cases you cannot hammer out.
>>>
>>>
>>> Best,
>>>
>>> Martynas
>>>
>>> On Sat, Dec 10, 2016 at 5:22 AM, Terry Roach <troach@capsi.com.au
>>> <mailto:troach@capsi.com.au>> wrote:
>>>
>>>    If I may interject in this debate, this all seems quite bewildering to me.
>>>
>>>    I am a pragmatic, practitioner of semantic technologies;  is a mere
>>>    consumer of W3C standards. Our company builds products based on your ideas
>>>    and so maybe I am not accustomed to how these things get cooked up, but
>>>    take a look at yourselves please. Somebody needs to inject a dose of
>>>    reality into this conversation.
>>>
>>>    We are very interested in the SHACL standard making it’s way through this
>>>    process and becoming endorsed so that we can commit to it in our products.
>>>    There will be no better test of the value and robustness of SHACL than the
>>>    community of semantic developers applying it in practice.
>>>
>>>    No standard is born perfect, of course it will evolve and I expect we will
>>>    find issues that will surely be addressed as it matures. But it needs to
>>>    get out of the door.
>>>
>>>    Perfection is the enemy of innovation here.
>>>
>>>    If there are any substantive issues with the standard, then of course
>>>    robust debate is great, but that should be in the form of a positive,
>>>    constructive suggestions. I am just seeing myopic, pedantic grandstanding
>>>    here.
>>>
>>>    There is a very vocal minority (of one) holding this debate hostage and it
>>>    is a travesty that the enormous effort that has gone into this piece of
>>>    work is being held up in this way.
>>>
>>>    Enough already
>>>
>>>
>>>    *Terry Roach*
>>>    Chief Executive Officer
>>>
>>>
>>>    Suite 105, International Business Centre, Australia Technology Park
>>>    2 Cornwallis St.
>>>    Eveleigh NSW 2015, Australia
>>>
>>>    M:  +61 421 054 804 <tel:%2B61%20421%20054%20804>
>>>    troach@capsi.com.au <mailto:troach@capsi.com.au>
>>>    www.capsi.com.au <http://www.capsi.com.au>
>>>
>>>
>>>
>>>
>>>
>
>

Received on Sunday, 11 December 2016 10:31:47 UTC