Re: HELP: Need At-Least-One model, and CDATA element type

> "Henry S. Thompson" wrote:
> 
> Nope, what you have below is (a, b?) | (a?, b)

Oops, misread your reply.  I took Jason's suggestion, which
was exactly what I wanted.

> I'm still not clear what you want.  Let's call your element
> 'brianBox'.  Which of the following do you want to be OK:
> 
> 1)
>    <brianBox><p>This is a <em>test</em> of the box</p></brianBox>
> 
> 2)
>    <brianBox><p>This is a <em>test</em> of the box</p>
>              <p>more stuff</p></brianBox>
> 
> 3)
>    <brianBox><p>this is not <xyzzy>HTML</xyzzy></p></brianBox>
> 
> 4)
>    <brianBox><p>this is not balanced</q></brianBox>

All of the above.  In some cases the contents of brianBox will be just
text.  In some cases it might contain markup of any kind (HTML, some
XML from another schema/dtd, etc.).  On other cases it might be
comma or tab separated tables, or even text encoded binary data
(GIF files, for example).  I basically want to accept just about
anything the UTF-8 encoding can handle.  The problem is, I don't
necessarily know or have any control over the contents of the
brianBox element.

Number 4 is the most frequent case, I believe.  Most of the time
the contents will be unbalanced HTML from older web documents,
etc.  Keep in mind, I don't want to validate or in any way examine
or handle the contents of brianBox.  I just want to transport it
within the enclosing document (which must conform to my schema).

Thanks!
Brian

-- 
Brian Atkins          brian_atkins@firehunter.com         (970) 288-2114
Agilent Technologies     4800 Wheaton Drive        Ft. Collins, CO 80525

Received on Tuesday, 11 September 2001 03:33:42 UTC