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

Re: A proposal for addressing LC comment 58 (fully typed functional-style syntax)

From: Christine Golbreich <cgolbrei@gmail.com>
Date: Fri, 20 Feb 2009 08:50:59 +0100
Message-ID: <b0ed1d660902192350u5fe355e7m33059caadd9e794a@mail.gmail.com>
To: Michael Schneider <schneid@fzi.de>
Cc: Boris Motik <boris.motik@comlab.ox.ac.uk>, W3C OWL Working Group <public-owl-wg@w3.org>
2009/2/19 Michael Schneider <schneid@fzi.de>:
> Hi!
>
> AFAICT, this would mean that many documents need a large revision:
>
> * Structural Spec
> * OWL/XML
> * Direct Semantics
> * RDF Mapping
> * Profiles
> * Test Cases
> * the UFDs that use Functional Syntax
>
> This would also mean that I would have to rework the proof of the
> correspondence theorem between OWL 2 DL and OWL 2 Full in the RDF-Based
> Semantics, and actually the whole (long!) Section 7 there. This may be a lot
> of work for me, that does not have any advantage for the RDF-Based
> Semantics. I would not be happy with this additional burden.
>
> We had a decision on this almost a year from now, after long email
> discussions, and no one seemed to be exceptionally unhappy with this. In
> particular, AFAIR (I might be wrong) Matthew was even a guest at F2F2. I'd
> say that the disadvantages on the implementer's side do not outweigh the
> additional burden on the (now small active part of the) WG.

Indeed. I even remember earlier discussions and opposition to fully
typed syntax when it was originally introduced.

Furthermore, while offering advantages for implementers, in
counterpart the requested change ignores a possible disavantage for
users:
Matt:
> (For example compare SubClassOf(A   ObjectSomeValuesFrom(R C)) with SubClassOf(A SomeValuesFrom(R C)).  Is
> R an object property or data property? With the strongly typed syntax  there is no ambiguity, but with the current syntax there is.

1) Dont' want to reiter old discussions, but this assertion seems to
rely on the assumption that it will be the user responsability to
select the right typed construct 'ObjectSomeValuesFrom' at top level
via an editor. Without user interaction, how can it be disambiguited
in a single pass (given that declarations are optional and punning
allowed) ?
Having to  decide  'some' or 'all' is already not always trivial for
users designing applications, having to select a 'type' (object/data)
in addition, makes it even more penalizing.  If it is decided to
impose the burden of fully typed syntax to users, would it not be then
cleaner from a specification viewpoint to rather impose mandatory
declarations ?

2) As asserted in other LC  responses, the functional syntax is aimed
at being clear and easy to read.  Also see e.g. an earlier email:
"the functional syntax should be in there to help people who are going
to read the other specs. Even a naive user might want to learn more
and having it available (though, not the default one) is helpful".
http://lists.w3.org/Archives/Public/public-owl-wg/2008Aug/0182.html
I'd recommand to keep that syntax as it, simple of use and readable by users.

3) Would it not imply to modify (user) syntaxes e.g. Manchester syntax
e.g. replacing Some by ObjectSome and DataSome ?
Does it assume multiple syntaxes ?

4) Finally, though it's obviously secondary, in first versions,
grammar and examples of NF&R were in fully typed syntax. Next I
updated to present syntax. I'll not really enjoy having to switch it
back again. But I can imagine it's even worse for other editors, e.g.
the Syntax, RDF etc. !

> I propose to deny the requested change.
>
> Michael

Agree.

Christine
>>-----Original Message-----
>>From: public-owl-wg-request@w3.org [mailto:public-owl-wg-request@w3.org]
>>On Behalf Of Boris Motik
>>Sent: Thursday, February 19, 2009 9:10 PM
>>To: 'W3C OWL Working Group'
>>Subject: A proposal for addressing LC comment 58 (fully typed
>>functional-style syntax)
>>
>>Hello,
>>
>>The Last Call comment 58 by Matthew Horridge asks us to change the
>>functional-style syntax back to the earlier state, where it was fully
>>typed.
>>Thus, instead of the current class expression
>>
>>SomeValuesFrom( a:P a:C )
>>
>>we would have the class expression
>>
>>ObjectSomeValuesFrom( a:P a:C ).
>>
>>This change would not affect in any way any of the other syntaxes,
>>OWL/RDF
>>included. The main rationale behind the change is to make functional-
>>style
>>syntax documents easier to parse.
>>
>>I actually sympathize with this comment. After the second F2F when we
>>decided
>>that we don't want strong typing in OWL/RDF, I rather arbitrarily
>>decided to
>>change the functional-style syntax as well. As Matthew's comment shows,
>>however,
>>this had rather unfortunate effects.
>>
>>We can address this comment by changing the functional-style syntax to
>>the
>>earlier state and make it fully typed. Since the primary syntax of OWL 2
>>is
>>OWL/RDF, and this syntax would not change in any way, I don't think that
>>this
>>would be a problematic change. Let me know how you feel about it. If
>>everyone
>>agrees, I shall draft a response to Matthew along these lines.
>>
>>Regards,
>>
>>       Boris
>>
>
> --
> Dipl.-Inform. Michael Schneider
> Research Scientist, Dept. Information Process Engineering (IPE)
> Tel  : +49-721-9654-726
> Fax  : +49-721-9654-727
> Email: schneid@fzi.de
> WWW  : http://www.fzi.de/ipe/eng/mitarbeiter.php?id=555
>
> ============================================================================
> ==
>
> FZI Forschungszentrum Informatik an der Universität Karlsruhe
> Haid-und-Neu-Str. 10-14, D-76131 Karlsruhe
> Tel.: +49-721-9654-0, Fax: +49-721-9654-959
> Stiftung des bürgerlichen Rechts
> Stiftung Az: 14-0563.1 Regierungspräsidium Karlsruhe
> Vorstand: Prof. Dr.-Ing. Rüdiger Dillmann, Dipl. Wi.-Ing. Michael Flor,
> Prof. Dr. rer. nat. Dr. h.c. Wolffried Stucky, Prof. Dr. rer. nat. Rudi
> Studer
> Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
>
> ============================================================================
> ==
>
>



-- 
Christine
Received on Friday, 20 February 2009 07:51:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 20 February 2009 07:51:38 GMT