Re: ACTION-215 Review Manchester Syntax Document

Hi Peter

I did not have time to review this version of the Manchester syntax,
but when I looked at its grammar into more details before, in some
previous versions, my impression was that the reverse mapping may even
raise some problems, because of the "frame-based" option which
underlies it. I must say that I don't remember now the exact points
where it was problematic (and I did not have sufficient reason to
investigate it better at this time). As you are involved in the
different syntaxes, you certainly know the best if it's only tricky,
but possible.

Also, the frame option of M. syntax does not seem consistent with *
12.1 Dropping the Frame-Like syntax * as asserted and explained in
details in that section. Besides, the frame-based display of  M.
syntax, perhaps influenced by the firstProtégé-frames tool, may be
misleading about OWL 2 which (unlike OWL 1) does not provide a
frame-like syntax.

This is one reason, among others, why I am not so enthusiastic about
having  the M. syntax presented as a syntax with a particular status,
though already being supported by different tools and having some
"user-friendly" aspect.

Regards

Christine

2008/10/8 Peter F. Patel-Schneider <pfps@research.bell-labs.com>:
>
> From: Michael Smith <msmith@clarkparsia.com>
> Subject: ACTION-215 Review Manchester Syntax Document
> Date: Tue, 07 Oct 2008 17:15:23 -0400
>
>>
>> I have completed my review of the Manchester Syntax document by making
>> some small edits and adding several review comments [1], dispatching
>> ACTION-215.
>>
>> In addition, the document specifies a mapping from Manchester syntax to
>> the functional syntax.  The mapping in the other direction, while
>> intuitive, is not specified.  I don't believe it a requirement, but
>> think the forward mapping would be nice to have.
>
> Hmm.  The forward mapping is going to be a bit tricky, not because of
> technical issues, but because the "best" mapping may not be completely
> obvious.  The easiest mapping to specify would be to have each axiom
> turn into a frame, but this is probably not the best mapping.  A much
> better mapping would be to gather "all" the information about a URL, for
> some definition of "all", and turn it into one frame.  I suppose that
> stating the easy mapping and also stating that tools may want to do some
> coalescing as long as the result corresponds to the original might be
> best.
>
> This would be a significant addition, and is not likely to get into the
> document before the next F2F.
>
>> --
>> Mike Smith
>>
>> Clark & Parsia
>>
>> [1] http://www.w3.org/2007/OWL/wiki/index.php?title=ManchesterSyntax&diff=13730&oldid=13384
>
> peter
>
>



-- 
Christine

Received on Wednesday, 8 October 2008 14:56:31 UTC