Re: interleave and strings

[James Clark <jjc@jclark.com>]

Yes.  Apart from the fact that TREX has anyString (which is like #PCDATA
in SGML and XML) instead of mixed, and that TREX has attributes in the
content model, the restriction is basically the same as XML Schema and
RELAX Core.

Another slight difference is that TREX tries to preserve the usual
axioms for <empty/> and <notAllowed/>.  Thus

<group>
  <data type="xsd:int"/>
  <empty/>
</group>

is allowed because it is the same as

  <data type="xsd:int"/>

and

  <group>
    <data type="xsd:int"/>
    <notAllowed/>
  </group>

is allowed because it is the same as

   <notAllowed/>


Murata Makoto wrote:
> 
> James Clark wrote:
> 
> > I think this
> > restriction is one that is natural and inoffensive to markup users, and
> > simple to grasp intuitively.  XML Schema has the same restriction.
> 
> To me, the definition of your restriction is difficult to understand.  But this
> sentence helps me to have some guess.
> 
> XML Schema and RELAX Core have the same restriction: when you declare some content
> model is "mixed",  not only whitespace characters but also non-whitespace characters
> are allowed to prepend, intersperse, and follow elements.  TREX is slightly more
> powerful in the sense that you can specify where in the content model <anyString>
> occur.  For example, (a, #anyString, b, c, #anyString) is possible in TREX, but
> XML Schema and RELAX Core only allows (#anyString, a, #anyString, b, #anyString,
> c, #anyString).  This is not quite it, since TREX allows <attribute ....> as part of
> content models.  But this is basically the difference.  Am I right?
> 
> Cheers,
> 
> Makoto