Re: [xml] Help with dtd, and XML_PARSE_PEDANTIC

[Daniel Veillard <veillard redhat com>]

On Tue, Dec 19, 2006 at 01:13:55PM +0000, chinlu chinawa wrote:
> Hello,
>
> I'm validating a xml document via xmlCtxtReadFile,
> with the flags XML_PARSE_DTDVALID |
> XML_PARSE_PEDANTIC.
>
> The problem I'm finding is with the structure of my
> document which looks like this:
>
> <question>
>   <title>bla bla</title>
>   <answer>bla bla</answer>
>   <correct-answer>bla bla</correct-answer>
>   <image>bla bla</image>
> </question>
>
> There can be more than one possible <answer>, but only
> one mandatory <correct-answer> per question, and zero
> or one <image>.
>
> Then I thought on letting the user being able to put
> those nodes in the order he/she felt like doing it,
> though I think this is not actually possible, is it?
>
> At dtd validation time "positional-strictness"
> applies, and there's no way I can do it (or I haven't
> achieved it, have tried many different dtd
> constructions for a while)
>
> So I wonder if someone can confirm me that I'm forced
> to provide a static structure so users would have to
> build the document strictly as dtd says is should
> (same order), or if there's someway I could have the
> document well validated while letting the users chuck
> these nodes in an arbitrary order?

  It's a limitation of DTDs, in XSD there is the xsd:all construct,
in RNG there is the interleave construct. With DTD you must specify 
explicitely all the orders possible e.g:

   (a, b, c) | (a, c, b) | (b, a, c) | (b, c, a) ...

 of course with 4-5 elements it's already a serious problem.
 Conclusion: DTD are extremely limited
 Also this has nothing to do with XML_PARSE_PEDANTIC, but with validation

Daniel

-- 
Red Hat Virtualization group http://redhat.com/virtualization/
Daniel Veillard      | virtualization library  http://libvirt.org/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine  http://rpmfind.net/