W3C home > Mailing lists > Public > public-owl-wg@w3.org > February 2009

Re: Proposal for use of labels in Manchester Syntax ISSUE-146, ACTION-247

From: Alan Ruttenberg <alanruttenberg@gmail.com>
Date: Wed, 4 Feb 2009 03:00:23 -0800
Message-ID: <29af5e2d0902040300hbf69453o70ca205ac41f384@mail.gmail.com>
To: Bijan Parsia <bparsia@cs.man.ac.uk>
Cc: Bijan Parsia <bparsia@cs.manchester.ac.uk>, W3C OWL Working Group <public-owl-wg@w3.org>

On Wed, Feb 4, 2009 at 2:50 AM, Bijan Parsia <bparsia@cs.man.ac.uk> wrote:
> On 3 Feb 2009, at 23:57, Alan Ruttenberg wrote:
>> 2009/2/3 Bijan Parsia <bparsia@cs.manchester.ac.uk>:
>>> On 3 Feb 2009, at 20:15, Alan Ruttenberg wrote:
>>> [snip]
>>>> That there is a choice of labels is due to a request by Alan Rector.
>>>> The motivation for a choice of language should be clear.
>>> Choice? I don't get this.
>> The labels: keyword says which annotation property can be used for a
>> quoted label. One might suggest that only rdfs:label be used for
>> labels but Alan said he wants to be able to use other properties, for
>> example skos:prefLabel
>> Similarly, one might want to be able to choose that labels used in
>> rendering to manchester syntax are ones in a particular language.
> This all seems to be tool level rather than language level. Why put this in
> the format?

I honestly don't see the point of the format without this.

> Swoop allowed one to configure the display based on lang tag. Easy to make
> it for property as well. Easy to share config files.
> It requires a bit more work for text editors, but it's not so hard.
>>> [snip]
>>>> Labels with highest precedence are associated with IRIs by being added
>>>> to LabelToIRI. The order of precedence is:
>>> [snip]
>>> My instinct right there is to say, "non starter".
>>> I have no idea why there's an order of precedence for this feature. It
>>> seems like a very, very bad idea.
>> Without a precedence ordering the mapping can be ambiguous. Even with
>> it, it can, but less often.
> Er...that makes me think that this is not a finished proposal. It seems to
> be a minimal requirement that a mapping be unambiguous in all cases.

The ambiguity isn't a function of the proposal. The ambiguity only
arises if the author of an ontology uses the same value for a label
property for two entities. One can't prevent the user from doing this.
In the case that the author attempts to use such a label as a
quotedLabel, the proposal defines the behavior (a syntax error). In
the case of serialization such labels would not be used.


> Cheers,
> Bijan.
Received on Wednesday, 4 February 2009 11:00:57 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:09 UTC