W3C home > Mailing lists > Public > public-owl-wg@w3.org > November 2007

Re: ISSUE-67 (reification): REPORTED: use of reification in mapping rules is unwise

From: Alan Ruttenberg <alanruttenberg@gmail.com>
Date: Fri, 23 Nov 2007 17:48:06 -0500
To: Jeremy Carroll <jjc@hpl.hp.com>
Message-Id: <BB3CE205-F3BA-4E79-8030-CF1BCA18095A@gmail.com>
Cc: Boris Motik <boris.motik@comlab.ox.ac.uk>, "'OWL Working Group WG'" <public-owl-wg@w3.org>

On Nov 23, 2007, at 8:22 AM, Jeremy Carroll wrote:

> While 'hard requirement' is a better term than 'need' I am not sure  
> that there are any such in an incremental release.

This is a fair criticism. Let me restate: Based on experience  
attending the OWLED F2F in the last few years on observing the  
genesis of what became the OWL 1.1 submission, in my assessment,  
annotations of Axioms is one of the more sought after improvements by  
people, such as Alan Rector, who engineer large systems based on OWL.  
Because it was requested by some of the people with the most  
experience using OWL, I had assumed that getting this in to OWL 1.1  
is a very high priority, particularly given the emphasis in the group  
on making OWL more useful to actual users. However, this being a  
different setting from the OWLED, we should get input from the rest  
of the working group as to whether it is a priority for them.

> OWL 1.1 will be as good as we can make it, within constraints.
> I am happy that we have high wants including:
> - annotation of axioms
> - a mapping of such annotations into RDF
> - an OWL Full semantics of such mappings
> - a change which is an incremental change over OWL 1.0, rather than  
> a major change
> I am currently far from convinced that all these desiderata can be  
> met simultaneously. It is arguably unhelpful then to describe any  
> of them as needs or even hard requirements. If we discover that  
> something has to give, then if all of these are indeed hard  
> requirements, we shall fail.

Fair enough. We should, however, be thinking about how we can  
prioritize wants so that we have some basis for making decisions,  
other than ruling out any change where all wants can be satisfied. It  
may indeed be the case that the WG decides to drop some issues based  
on such considerations, but I would worry that dropping annotations  
on Axioms at this point, given the expressed need, is premature, and  
would rather see more creative work on how to make something like  
this work well and satisfy all or most wants.


> Jeremy
> Alan Ruttenberg wrote:
>> On Nov 21, 2007, at 6:40 AM, Boris Motik wrote:
>>> Hello,
>>> I don't want to get into an argument here whether this is really  
>>> needed or not; however, I wanted to point out that I spoke to quite
>>> a few people asking for the annotation of axioms.
>> I concur that this is a hard requirement for OWL 1.1. It has come  
>> up over and over again at OWLED.
>>> There is yet another solution: we might have axiom annotations in  
>>> the structural specification, but then disallow (or simply delete
>>> them) when the ontology is exported into OWL RDF.
>>  From my point of view, if some feature can't render in RDF, then  
>> they might as well not be defined at all. Having a complete RDF  
>> rendering of the full language would seem to me to also be a hard  
>> requirement.
>> Regards,
>> Alan
>>>     Boris
>>>> -----Original Message-----
>>>> From: public-owl-wg-request@w3.org [mailto:public-owl-wg- 
>>>> request@w3.org] On Behalf Of Jeremy Carroll
>>>> Sent: 21 November 2007 11:37
>>>> To: Boris Motik
>>>> Cc: 'OWL Working Group WG'
>>>> Subject: Re: ISSUE-67 (reification): REPORTED: use of  
>>>> reification in mapping rules is unwise
>>>> Boris Motik wrote:
>>>>> Hello,
>>>>> I actually really dislike reification myself; unfortunately, I  
>>>>> don't see how to get around these
>>>> issues in certain cases. The
>>>>> problem is that sometimes you need more than binary  
>>>>> associations between objects.
>>>>> For example, consider the problem of annotating a SubClassOf  
>>>>> axiom. In RDF, you write <x
>>>> rdfs:subClassof y>. But you've just used
>>>>> both x and y; there is no place for an annotation.
>>>>> The only solution I see is not to use reification, but to  
>>>>> introduce yet separate vocabulary and
>>>> represent ternary relations more
>>>>> explicitly. I am really open to any suggestions on this point,  
>>>>> because I do see the point that
>>>> reification is ugly.
>>>>> Boris
>>>> A different approach would be to decide that we cannot address  
>>>> the use
>>>> cases for annotations of axioms yet, and to postpone related  
>>>> issues, and
>>>> make do with the OWL 1.0 solution.
>>>> Jeremy
Received on Saturday, 24 November 2007 15:33:25 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:00 UTC