Re: [relax-ng] Algebraic characterization of RELAX NG patterns

James Clark wrote:

> Thanks for taking the time to explain this.  I found it quite hard to
> understand this.  I *think* I do now understand it.  Let me check my
> understanding by attempting a more informal, intuitive explanation.
> Hopefully, Murata-san will correct me if I have misunderstood.

Your understanding is correct.

James Clark wrote:

> 2'. The number of occurrences isn't significant (i.e. if you have one
> attribute from a particular name class, having another attribute from
> the same name class won't affect the match)

Yes, this is the point of the restriction.  If a name class is actually 
a single name, XML 1.0 allows only one attribute of this name class.  Thus, 
2' holds.

Cheers,

Makoto