OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   RE: [xml-dev] SGML queries

[ Lists Home | Date Index | Thread Index ]

That's one.  I'll agree that if one is doing a 
lot of conversion work, it might be best to slip 
back to SGML and use the power tools.  The example 
you give is interesting in terms of a production 
flow in and out of DTDs where each DTD describes a 
state in the production process.  In the web 
industry were better informed, SGML expertise would 
be one of the things to check when discovering a 
company and considering it for conversion work.

I have done "ugly" conversions, but we did not 
use that technique.  I admit to thinking more 
in terms of documents with longer lifecycles 
rather than short, stepped versions.  I think 
the handbook got it right though in saying that 
techniques such as a fully minimized element 
are for those who completely understand the 
contexts of the declarations. Otherwise, as 
it also says, valid SGML can be produced that 
does not match the intent of the designer. 

Since we often design for people who barely 
bother to understand what the brackets are for, 
using a fully minimized element seems risky 
unless one has full control of all the places, 
people, and processes in which it is used. 
For a web DTD shared outside a closed system, 
it seems even riskier.  OTOH, the HTMLers 
used it.  It may have worked simply because 
the tag stackers ignored the DTD in most cases, 
and when they did bother to validate, the DTD 
proves so loose as to almost be worthless. 

That's an interesting rabbit trail in and 
of itself:  how useful has HTML validation 
and the HTML DTD proven to be other than being 
a formal description of the tags in the soup?


From: Marcus Carr [mailto:mcarr@allette.com.au]

Bullard, Claude L (Len) wrote:

> *Bonehead elements* might be good.  Other than in 
> the SGML Handbook, I've never seen these used 
> in practice.

You haven't spent enough time around ugly (read "cost effective by 
coding alone") markup jobs then. I've used omissable start and end tags 
more times than I could count - the classic situation is for turning 
something like:

   <section>Laundry basket
   <para>The laundry basket has a long and fascinating history...

into something like:

   <title>Laundry basket</title>
   <para>The laundry basket has a long and fascinating history...

It's intuitive, equivalent and less markup. I'm surprised that there's 
so little enthusiasm for this - I've had markup teams ask for this sort 
of change to be added to a DTD. Omission is one of the key drivers that 
has kept us using SGML en route to the delivery of XML data. Normalizers 
are cheap, markup people aren't.


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS