Lists Home |
Date Index |
I have not checked it for quite some time. Efficiency is one issue,
however 1-unambiguity gets in the way as such -- it is an open problem to
obtain the strictest possible 1-unambiguous grammar given an operation and
a 1-unambiguous grammer. If we allow 1-ambiguity, with or without
ambiguity in interpretation (I hope you remember Makoto's talk at Extreme
last year regarding interpretation - it is basically the same as type
assignment), it becomes much easier.
Anyways, I have not studied the claims of efficiency due to 1-unambiguity
-- I will check them out some time. I have to do my homework.
Also, we are talking about efficiency of schema operations -- and schema
operations usually take a much smaller time when compared to the
operations on the data. So does it really matter whether something is
exponential or linear in the size of the schema??
regards - murali.
On Wed, 20 Mar 2002, Jonathan Robie wrote:
> At 12:06 PM 3/20/2002 -0800, Murali Mani wrote:
> >The biggest problem with XML Schema spec as I see is the 1-unambiguity
> >constraint -- this gets very badly in the way of XML Query, as I
> >understand it. I wonder what others feel about it.
> Actually, 1-unambiguity has real advantages for efficient implementation of
> structural subsumption. I don't think it particularly gets in the way of
> XML Query. I used to think that it would.