Re: PSVI

[Sean McGrath <sean@digitome.com>]

At 07:19 AM 3/3/01 +1100, Justin Couch wrote:
[...]
>We
>really need the DTD or schema to be processed along with our document to
>build all of the correct element information for rendering. We can't
>just rely on having the software produce correct values for us.  Even
>with a non-validating parser, we need this information from somewhere.
>Having the user-land code assume that it knows the default values
>doesn't really make life particularly easy for us for consistent
>behaviours. So we're a case of must have DTD and/or Schema even when we
>don't need them as per spec.

Let me rephrase that and see if you agree or dis-aggree. You need
a transformation (adding attribute values to elements) to occur
after validation. The only "one ring pull" options you have
are DTD or Schema. You could do it with XSLT or some
other transformation system but it would involve (as things
stand now) two ring pulls?