Re: Fwd: Re: Fwd: On complex constraints

Sure!

When I stated in Lisboa it was not possible I was only meaning /with the 
current OWL specifications/.
I was unaware, though, of the W3C Note with the proposal and the 
RacerPro implementation.

Very instructive references! Thanks a lot!
Víctor

El 07/11/2016 a las 13:33, Sabrina Kirrane escribió:
> Dear all,
>
> Please see below response from Piero regarding representing complex
> constraints in DL.
>
> Best Regards,
> Sabrina
>
>
> -------- Forwarded Message --------
> Subject: Re: Fwd: On complex constraints
> Date: Thu, 3 Nov 2016 11:26:05 +0100
> From: Piero Bonatti <pieroandrea.bonatti@unina.it>
> To: Sabrina Kirrane <sabrina.kirrane@wu.ac.at>, Axel Polleres
> <axel.polleres@wu.ac.at>
> CC: Simon Steyskal <simon.steyskal@wu.ac.at>
>
> Hi everybody,
>
> In DL one can use concrete domains, in principle. It would look more or
> less like this:
>
> declare that Photos have:
>   a property "about" whose range includes football games;
>   a concrete feature "hasPublicationDate" with datatype Date.
>
> In turn football games have a concrete feature "hasEndTime".
>
> then assert something like:
> Subclass( PublishablePhoto
>            (>= hasPublicationDate about.hasEndTime+1))
>
> (note: operator '+' might not be directly supported in that context, but
> one can overcome this limitation using an auxiliary feature to hold the
> result of the sum)
>
> OWL 2 does not (yet) support linear constraints like the above. However:
>
> 1) There are proposals to support them, see
> https://www.w3.org/TR/2012/NOTE-owl2-dr-linear-20121211/
>
> 2) The theory is all there (syntax, semantics, decidability), see
> http://sunsite.informatik.rwth-aachen.de/Publications/CEUR-WS//Vol-86/13-Lutz.ps
>
> 3) The RacerPro engine already supports linear constraints, see
> http://www.dblab.ntua.gr/~gtsat/collection/RDF%20stores/Racer_whitepaper.pdf
>
> there might also be more engines supporting linear constraints, but I
> have no time to look for them right now
>
> cheers
> Piero
>
> PS: the syntax I used above is not exactly the syntax used in 1), nor
> the syntax used on 3), so if you need to show something formally correct
> then you'll have to adapt my encoding accordingly.
>
> On 11/10/2016 00:04, Sabrina Kirrane wrote:
>> Hi Axel & Piero,
>>
>> On this weeks W3C POE call we were discussing how we can add complex
>> constraints (i.e. constraints on constraints) to ODRL.
>>
>> At the moment it is possible to associate 0:M duties with a permission.
>> Duties can have 0:M constraints, and each one must be satisfied in order
>> to gain access to the resource (N.B OR and XOR operators are currently
>> not supported, however there is talk about reintroducing them).
>>
>> Going beyond simple boolean operators TR have indicated that there is a
>> need to express relations between constraints.
>>
>> For example they need to be able to express relative events e.g. a
>> relative constraint could state a photo may be published "1 hour after
>> the football game ended".
>>
>> How would you express this in RDF? OWL?
>>
>> Bet regards,
>> Sabrina
>>
>>
>> -------- Forwarded Message --------
>> Subject: 	On complex constraints
>> Resent-Date: 	Fri, 23 Sep 2016 09:37:59 +0000
>> Resent-From: 	public-poe-wg@w3.org
>> Date: 	Fri, 23 Sep 2016 11:43:20 +0200
>> From: 	Víctor Rodríguez Doncel <vrodriguez@fi.upm.es>
>> To: 	public-poe-wg@w3.org
>>
>>
>>
>> Dear all,
>>
>> I did not verbally express my view as I did not feel capable to. I'll
>> try to write it:
>>
>> Simple constraints are in ODRL logical expressions, whose truth value
>> might be evaluated. With ODRL we can express for example:
>>
>> R1: industry = academy
>> R2: count <=5
>>
>> We have in ODRL the ability to define logical expressions with two
>> variables: an "operator", a "rightOperand" and a left operand (called
>> "name").
>> We can even combine these expressions with connectives (extended model):
>> AND, OR, NOT.
>> However, we are moving within the limits of logic and we cannot evaluate
>> algebraic expressions.
>> Namely, we would like to have:
>>
>> R3: publicationTime > (asset.FootballMatchEndTime + 1 hour)
>>
>> We cannot define algebraic formulas, nor variables. We would like to
>> have something like:
>>
>> var1: asset.FootbalMatchEndTime + 1 hour
>> R4: publicationTime > var1
>>
>> So even if we use reification, we can't still represent the addition
>> operation.
>> I think adding these features would dramatically increase the complexity
>> of the model...
>>
>> Víctor
>>
>> El 23/09/2016 a las 9:49, Renato Iannella escribió:
>>> Raw minutes from
>>> yesterday: https://www.w3.org/2016/09/22-poe-minutes.html
>>>
>>> Renato Iannella, Monegraph
>>> Co-Chair, W3C Permissions & Obligations Expression (POE) Working Group
>>>
>>

-- 
Víctor Rodríguez-Doncel
D3205 - Ontology Engineering Group (OEG)
Departamento de Inteligencia Artificial
ETS de Ingenieros Informáticos
Universidad Politécnica de Madrid

Campus de Montegancedo s/n
Boadilla del Monte-28660 Madrid, Spain
Tel. (+34) 91336 3753
Skype: vroddon3