RE: SISR: Comments on SISR April 2003 working draft

Sturtevant Dean wrote:

>
> I have a comment on one of Mr. Wyss's proposals. Note I am not
> a member of any W3C committee:
>
> He said:
> >  3) Clarification of GRR of external rule references
> >
> > Section 3.3.2 does not specify how the GRN of external rule
> > references can be accessed in the referencing rule.  I suggest
> > that the GRN of external rule references is only available as
> > $$ and its text variable as $$$.text in the referencing rule.
> > Thus, there is no equivalence to "$rulename" for external
> > rule references.  If the value of the GRR is required past the
> > next rule reference, it has to be assigned to a temporary
> > variable.  
> > An alternate solution would be to extend the syntax of external
> > rule references to allow specifying an alias for the GRR. 
> > For example: 
> >
> >  $rule =
> >    $<http://example.com/foo.gram>^<$foo> 
> >    $<http://example.com/bar.gram>^<$bar> 
> >    {$.result = $foo + $bar};
>
> I respond:
> There's no need for new syntax here. It seems that the following
> should work (as is implemented on our platform):
>  $rule =
>    $<http://example.com/foo.gram#foo>
>    $<http://example.com/bar.gram#bar>
>    {$.result = $foo + $bar};

Regarding Dean's answer, it may be some kind of workaround, but is
nothing near a general solution.
What if an external rule reference is done without mentioning the
rule name (because the referenced module has a root rule)?
What about a rule like the following?
  $rule =
    $<http://example.com/foo.gram#r1>
    $<http://example.com/bar.gram#r1>
    {$.result = $r1 + $r1};
                ^^^   ^^^



Regarding Felix Wyss's post, I'm glad to see that problems/omissions
in the SISR standard are generating more posts!
I posted two messages concerning SISR on www-voice in the last 6
months (10th June 2003, 30th July 2003), but there hasn't been
any reaction on this list.
This surprised and still surprises me, since I would imagine that
contributors to said document are subcribers of this mailing list...

Does someone know what the current status of SISR is?
Is there any revision process planned?


--Serge Le Huitouze

Received on Tuesday, 2 December 2003 09:04:00 UTC