Re: Name classes

[Kohsuke KAWAGUCHI <kohsuke.kawaguchi@eng.sun.com>]

I hope OASIS would enhance the mailing list capability and substitute
"Reply-To" field by the address of ML.


> You are right, but the syntax is awkward.

Maybe it's time for <concur> to revive for name class!  (half seriously)


And as James said before, we can always check whether NC accepts
something or not.  I tried to extend the algorithm to compute the
cardinality of NC (finite or infinite), but it's probably impossible due
to <difference>.


----------------------- Original Message -----------------------
 From:    Murata Makoto <mura034@attglobal.net>
 To:      Kohsuke KAWAGUCHI <kohsuke.kawaguchi@eng.sun.com>
 Date:    Sat, 05 May 2001 03:01:41 +0900
 Subject: Re: Name classes
----

Kohsuke KAWAGUCHI wrote:

> 
> > 6) <intersection> on top of 3)
> 
> Intersection can be expressed in the current syntax.
> 
> <not> NC </not>  =>  <difference> <anyName/> NC </difference>
> <intersection> NC1 NC2 </intersection>
>   => <difference> NC1 <not>NC2</not> </difference>
> 
> So current name class is closed to boolean operations
> (union/intersection/difference). Am I right?

You are right, but the syntax is awkward.

Cheers,

Makoto
--------------------- Original Message Ends --------------------

--
Kohsuke KAWAGUCHI                          +1 650 786 0721
Sun Microsystems                   kohsuke.kawaguchi@eng.sun.com