Re: AFs and the DPH

Rick Jelliffe (ricko@allette.com.au)
Fri, 3 Oct 1997 02:35:37 +1000



> From: W. Eliot Kimber <eliot@isogen.com>

> This is why I think architectures are key to
> the success of XML: it lets you eat the cake of DTD-less documents and
> still have it (because the architecture processing gives you all the
> validation and processing you need, but only when you want it and not
> when you don't).

This seems a very good and important point.

If the problem is how to represent occassional structures in well-formed
documents, then AFs represent an external form, XML-data represents an
inline form, and ISO 8879 declarations represent a header form.

But, I think that a document with AFs cannot be regarded as being
declaration-less, since either the declarations have to be implicitly
built into the application, or be explicit in the form of a DTD outside.

The horrible thing is that, of course, there is no reason why an
XML-data schema could not itself be a meta-DTD! I think the
issue of direct modelling (SGML templates or XML-data) versus
indirect modelling (AFs) should be distinguished from the issue
of the goodness of ISO 8879 declaration syntax versus XML-data
non-standard syntax. AFs, as a mechanism, are syntax-neutral
to a great extent.

Rick Jelliffe