various issues

Peter Murray-Rust Peter at
Thu Apr 10 01:55:52 BST 1997

In message <v03007800af6fb40da9ef@[]> Neil Bradley writes:
> 2. Do others agree that there should be recommended models for representing
> complex structures such as tables and mathematical formulae, even if this
> just means referring to an existing standard, such as CALS tables or HTML
> tables, and ISO 12083 math? At this early stage it would be great if
> potential XML software developers had a common goal, that would ultimately
> benefit users of XML.
I am extremely keen on this idea and have suggested that we have 'Generally
Agreed Conventions'.  [The WG/ERB has rightly agreed that *it* doesn't have 
the resources and perhpas isn't the right place anyway.]  What I *don't*
want to happen is that everyone sits down and hacks a set of XML tags and fires
off WF documents at random (especially if there is no processing software
or 'to read this document you must have installed FOO.exe on an XYZ platform
running BAZ with PQR libraries...  I've seen a horrible example of this in
a talk today...).  So we should promote the *existence* of existing DTDs
and explain what they can offer.  So, for example, it is foolish for me to
hack CML so that it includes FIGURES - much better to interoperate with CGM
(recently mentioned on c.t.s).  I would hope that any chemical applications
would consider using some or all of the CML tagset.  And when parsable maths
comes along I will certainly adopt that.

The major problem is namespace collisions between ElementTypes.  Does anyone
have a solution?  What if the incompatible document is in a separate valid
document (i.e. with DOCTYPE).  I'd value ideas/epxerience here, otherwise
we will have to produce some kludge.

If we can get this off the ground before too many uncoordinated hacks appear
we have a chance to reduce the worst namespace pollution.  Yes, I realise that
formally every XML application is completely distinct from every other, but
it won't always work that way.  XML is EXTENSIBLE, so people will want to
combine DTDs or parts of DTDs...

It would be EXTREMELY valuable if someone were to offer to run a Generally
Accepted Conventions page.  This would include some basic guidelines about
what makes agood XML document and what doesn't, some commonly used tag sets,
advice on using (say) IDs, etc.


Peter Murray-Rust, domestic net connection
Virtual School of Molecular Sciences

xml-dev: A list for W3C XML Developers
Archived as:
To unsubscribe, send to majordomo at the following message;
unsubscribe xml-dev
List coordinator, Henry Rzepa (rzepa at

More information about the Xml-dev mailing list