Re: Is XML 1.0 underspecified? (was: Re: CDATA by any othername...)

Paul Prescod (papresco@technologist.com)
Sat, 31 Oct 1998 07:01:48 -0600


Sean Mc Grath wrote:
>
> What he said. If the W3C had landed the XML 1.0 in Z or VDM or
> something we wouldn't have half as many implementations as we
> currently have. Sure we now need more formalism to ensure that
> XML goes from strength to strength but without the balance of
> approachability/formalism XML 1.0 uses we would not have got
> here.

Given sufficient technical resources, you can have a very terse, precise
specification and tutorials and "annotated specifications" that provide an
introductory path to the specification. In the long run, reading a book on
Z, a book on XML and a Z-based XML spec is easier than trying to
reconstruct the ideas in the authors heads based on imprecise prose. It's
not like making an XML parser is a weekend job anyhow!

I'm not arguing in favour of Z in particular, however. I'm arguing in
favour of some form of precise formalism. One approach might be to come up
with a vague mental model, build non-normative formal descriptions to make
sure that the ideas are sound and complete, and then transliterate into
English.

> It must also be rememembered that SGML from which XML sprung
> has very complex interplays between parsing modes, logical and
> physical structures. Some of this was bound to leak over into
> XML.

That is certainly true. If we could go back in time and specify SGML in
terms of formalisms, it wouldn't be such a hairy language today. Even the
relatively simple grove formalism has lain bare the weirdness in SGML (but
also highlighted some deep and beautiful symmetries).

Hind sight is 20/20.

Paul Prescod - http://itrc.uwaterloo.ca/~papresco

"The new revolutionaries believe the time has come for an aggressive
move against our oppressors. We have established a solid beachhead
on Friday. We now intend to fight vigorously for 'casual Thursdays.'
-- who says America's revolutionary spirit is dead?