RE: Incompatibilities in RDQL implementations

That's a good idea - the member submission.  Until DAWG publishes.

The other one you referred to is the exact output of jjdoc so it is right
(as in eactly what my impl uses) but is more complex as a result and has
some overhead in the way it does things.  And it has no terminals in it as
jjdoc does not produce that output.

The W3C version [1] does not have the backward compatibility features but
otherwise should be OK.

There a bug in that grammar : <SUCHTAHT> is the same as <AND> and is the
token 'AND'.  I produced a special simplief version with javacc but had to
hand edit the final result.  I got tokenm blindness after while, I think.

I have asked DanBri if it is easy to do an errata.

	Andy

http://www.w3.org/Submission/RDQL/ 

-------- Original Message --------
> From: Arjohn Kampman <mailto:arjohn.kampman@aduna.biz>
> Date: 6 May 2004 14:38
> 
> Arjohn Kampman wrote:
> > Hi Andy, others,
> > 
> > Is the RDQL grammar at [1] up-to-date? If not, could you consider
> > updating it? Also, could you consider defining the terminals like
> > <SUCHTHAT>, <STR_MATCH> and <STRING_LITERAL1> in the grammar for the
> > sake of making it complete and to prevent deviations in
> > implementations? Currently, one has be dive into the code to find out
> > the definition of the terminals.
> 
> Never mind, Jeen just pointed me at the grammar in the W3C submission.
> 
> Arjohn

Received on Thursday, 6 May 2004 10:21:46 UTC