Re: Role of SPARQL

My personal opinion is that as a rule language, we should take something as simple as a restricted form of:

 { construct-template } IF { SPARQL-pattern}

Note: 

1) I think that we need a restricted form because there are potential non-termination issues

2) this is both simpler, but also more powerful (e.g. allowing SPARQL-aggregates in bodies) than SWRL, particularly, if you combine it with SPARQL's entailment regimes.

3) SPARQL is a standard. SWRL is not a W3C standard, but "only" a memeber submission.

4) we could also refer to RIF, which *IS* a standard, but has no readable syntax... so this is better, IMO. Also, note that RIF is not ideal as a basis, since it has datatype reasoning on board already in RIF Core, which is what you don't always want.

HTH, best,
Axel

--
Prof. Dr. Axel Polleres
Institute for Information Business, WU Vienna
url: http://www.polleres.net/  twitter: @AxelPolleres

On 25 Nov 2014, at 22:12, Peter F. Patel-Schneider <pfpschneider@gmail.com> wrote:

> Yes, there would have to be some work done.   However, there already is a semantics for SWRL in the W3C SWRL submission http://www.w3.org/Submission/SWRL/ so I think that the bulk of the work is done for OWL+SWRL.
> 
> peter
> 
> 
> On 11/25/2014 12:50 PM, Irene Polikoff wrote:
>> Yes, without a doubt - defining semantics will be a work item and a deliverable.
>> 
>> What I meant by 'requiring work' (and I have used your words) is that before the semantics of constraints could be defined one would need to do some work on extending/enhancing/modifying (whatever is more precise) semantics of OWL or RDF. So, there would be two pieces of work.
>> 
>> Please correct if I misunderstood you.
>> 
>> Irene
>> 
>> -----Original Message-----
>> From: Peter F. Patel-Schneider [mailto:pfpschneider@gmail.com]
>> Sent: Tuesday, November 25, 2014 3:26 PM
>> To: Irene Polikoff; 'Dean Allemang'; 'Holger Knublauch'
>> Cc: public-data-shapes-wg@w3.org
>> Subject: Re: Role of SPARQL
>> 
>> I would say that each solution for defining semantics requires work, unless the solution is to use SPARQL itself as the complete solution.  No proposal that I have seen is like this.
>> 
>> peter
>> 
>> 
>> On 11/25/2014 11:27 AM, Irene Polikoff wrote:
>>> < Perhaps there is something wrong with the SPARQL algebra that needs
>>> to be fixed so a parallel solution has to be developed.>
>>> 
>>> Peter, are you already aware of anything wrong with SPARQL algebra or
>>> is this a plan B in case it is discovered that there is something
>>> wrong with SPARQL algebra?
>>> 
>>> So, the options for defining semantics of constraints so far are:
>>> 
>>> 1.SPARQL
>>> 
>>> 2.OWL+SWRL semantics - would require work on adding features of SWRL -
>>> either SWRL itself or expressions that use SWRL functions. And, I
>>> suppose, would require defining the new, closed word OWL semantics.
>>> 
>>> 3.RDF semantics - would require work similar to the one above for
>>> using OWL semantics
>>> 
>>> 4.Algebra on RDF graphs and datasets - an alternative to using SPARQL
>>> in case there is something wrong with the SPARQL algebra that needs to
>>> be fixed
>>> 
>>> 5.Z semantics
>>> 
>>> Does anyone have another option they are wanting to be considered?
>>> 
>>> Irene
>>> 
>>> -----Original Message-----
>>> From: Peter F. Patel-Schneider [mailto:pfpschneider@gmail.com]
>>> Sent: Tuesday, November 25, 2014 8:43 AM
>>> To: Irene Polikoff; 'Dean Allemang'; 'Holger Knublauch'
>>> Cc: public-data-shapes-wg@w3.org
>>> Subject: Re: Role of SPARQL
>>> 
>>> One option for extending the constraint power of an OWL solution would
>>> be to add some features from SWRL.  This could either be SWRL itself
>>> or expressions that use SWRL functions.
>>> 
>>> I wasn't advocating the use of Z, just pointing out that it could be an option.
>>> 
>>> Basing a solution on the RDF semantics would require work similar to a
>>> solution based on the OWL semantics.
>>> 
>>> A solution using an algebra on RDF graphs and datasets might look very
>>> much like SPARQL.  Perhaps there is something wrong with the SPARQL
>>> algebra that needs to be fixed so a parallel solution has to be developed.
>>> 
>>> My email wasn't advocating any particular position, just pointing out
>>> that there are potential alternatives to SPARQL.
>>> 
>>> peter
>>> 
>>> On 11/24/2014 04:20 PM, Irene Polikoff wrote:
>>> 
>>>  > Dean,
>>> 
>>>  >
>>> 
>>>  > Your messages are indeed getting through.
>>> 
>>>  >
>>> 
>>>  > It seems to me that another issue with using OWL to do the kind of
>>> 
>>>  > definitions you are describing is that it can’t (to my knowledge)
>>> 
>>>  > cover a set a fairly common constraints such as start date must be
>>> 
>>>  > before the end date. I presuming here that this category of
>>> 
>>>  > constraints is accepted as a requirement. I believe Peter suggested
>>> 
>>>  > addressing this issue by using SWRL, so this option would be
>>> 
>>>  > OWL+SWRL. Is this correct?
>>> 
>>>  >
>>> 
>>>  > Peter identified a couple of other options:
>>> 
>>>  >
>>> 
>>>  > ·Z – I don’t think this is a viable idea as it introduces a new
>>> 
>>>  > language when there are already good options within the RDF stack
>>> 
>>>  >
>>> 
>>>  > ·RDF semantics – can this work? And how?
>>> 
>>>  >
>>> 
>>>  > ·Algebra on RDF graphs and datasets – can this work? And how?
>>> 
>>>  >
>>> 
>>>  > Irene
>>> 
>>>  >
>>> 
>> 
>

Received on Tuesday, 25 November 2014 21:20:51 UTC