Re: PS grammar question

On Sun, 21 Sep 2008 13:45:05 -0700
"Hassan Ait-Kaci" <hak@ilog.com> wrote:

> Hi Michael,
> 
> I agree with you as far as making the PS into "yet another rule language" -

Hassan, it wasn't me who was objecting :-)

> but it must as least be parsable, don't you think?

Yes. And this means "yet another language" in my understanding.

>  Besides, my effort here
> is only meant to try and work out a rapid prototype that would ease the
> automatic generation of XML from (A)PS in order to make the examples of
> RIF documents easy to manage (thus avoiding mistakes of the kind of the
> typos in the examples that I reported).

As I said, I don't object. My only concern is that we should not make it too
complicated. Throwing a few delimiters here and there should make it
parsable with reasonably little change.

This is not the only problem that exists in PS due to gratuitous objections
early on. I see couple of others, which I would be happy to get rid of:

1. The requirement that each symbol appears in a unique context.
   Not only this constraint has no good reason to exist, it creates
   difficulties in DTB when we want to define polyadic builtins.

2. The requirement that every variable must be *explicitly* quantified.
   This just makes the test cases hard to read and is totally unnecessary.
   

michael

 
> Without some form of delimiters like double quotes around IRI's in the
> BNF rules for Base and Prefix (which only appear there in "raw" form),
> the lexer would become context-sensitive (i.e., the nature of tokens
> being read depending on what is being parsed). With this simple innocuous
> change in the PS BNF, the problem disappears. I hope we can resolve this
> small issue at the next conf call.
> 
> -hak
> --
> Hassan Aït-Kaci  *  ILOG, Inc. - Product Division R&D
> http://koala.ilog.fr/wiki/bin/view/Main/HassanAitKaci
> 
> 
> 
> -----Original Message-----
> From: Michael Kifer [mailto:kifer@cs.sunysb.edu]
> Sent: Sun 9/21/2008 10:19 PM
> To: Hassan Ait-Kaci
> Cc: harold boley; RIF WG
> Subject: Re: PS grammar question
>  
> 
> 
> On Sat, 20 Sep 2008 18:15:25 -0700
> "Hassan Ait-Kaci" <hak@ilog.com> wrote:
> 
> > > Looks like a typo. Should be <http:...> or the full "..."^^rif:iri
> > 
> > I surmised so, but wanted to check. Harold should confirm or explain.
> > 
> > What about the following remark (which is really causing me a problem):
> > 
> > > PS/ BTW, the PS grammar's tokenizing is now complexified due to using PS
> > >     to declare Prefix and Base pragmas not using double-quoted strings
> > >     around the IRI's. The alternative would be to parse IRI's - which is
> > >     beyond such a prototype's goal. In the canonical PS, all such IRI's
> > >     are double-quoted strings which greatly simplifies the tokenizing.
> > >     It'd be as simple and as easy to do so for the Prefix and Base pragmas.
> > 
> > Can we change the PS syntax to require double-quotes around the IRI's
> > in the pragmas Prefix and Base? (I call them pragmas because all they do
> > is declare the IRI name spaces used in CURIE shorthands.) It would be a
> > simple amendment to the PS BNF and would immensely simplify implementing
> > a lexer for the PS.
> 
> The PS syntax is abstract (for the most part). It does not specify the
> delimiters and such (because some people were adamantly against "yet another
> language"). 
> 
> We could either add delimiters to PS and make it into a parsable language
> or do it separately. The first route would probably be preferable if the old
> objections won't be raised again.
> 
> michael
> 
> 
> > Thanks.
> > 
> > -hak
> > --
> > Hassan Aït-Kaci  *  ILOG, Inc. - Product Division R&D
> > http://koala.ilog.fr/wiki/bin/view/Main/HassanAitKaci
> > 
> > 
> > 
> > -----Original Message-----
> > From: Michael Kifer [mailto:kifer@cs.sunysb.edu]
> > Sent: Sun 9/21/2008 2:40 AM
> > To: Hassan Ait-Kaci
> > Cc: harold boley; RIF WG
> > Subject: Re: PS grammar question
> >  
> > 
> > 
> > On Sat, 20 Sep 2008 17:32:51 -0700
> > "Hassan Ait-Kaci" <hak@ilog.com> wrote:
> > 
> > > > (* "http://sample.org"^^rif:iri pd[dc:publisher -> http://www.w3.org/
> > > >                                    dc:date -> "2008-04-04"^^xs:date] *)  
> > > 
> > > Can you please explain to me how http://www.w3.org/ can be derived as a
> > > TERM according to the EBNF rules cited above?
> > 
> > Looks like a typo. Should be <http:...> or the full "..."^^rif:iri
> > 
> > m
> > 
> 

Received on Sunday, 21 September 2008 21:04:33 UTC