FW: Namespaces and DTDs

[Clark Evans]

XML is really nasty for humans to look at.  SGML is 
definately better.  SGML is really nasty for computers
to look at XML is definately better.

Didier PH Martin wrote:
> 
> The whole point here is: if I need validation at the receiving end why 
> not use SGML which is more elaborate and necessarily need validation 
> (because of the > possibility to have omittags). If however, we do not 
> need validation at the receiving end then, we are better to use XML that, 
> because of its structure, can be parsed without validation and then a DOM 
> could be created for procedural language consumption.

I see it this way:

	XML = { SGML - irregular grammar constructs }

Thus, in general, you can't parse an SGML document without
the DTD describing it's grammer, where you can parse 
an XML without the DTD.  

An SGML DTD has two components:  

	a) Those parts used as grammer
	b) Those parts dictating structure

Where an XML DTD only has the latter (structural declarations)

Why is this important?

I was visiting a friend of mine with a GUI tool (Netscape 
Composer?) trying to write a web site.  She was very frustrated
beacuse she could not get the page to do what she wanted.
For my next visit, I bought her a copy of the HTML O'Rilley 
book.  I sat down with her and showed her how to write HTML
manually...  it was like pulling teeth.  But then, after about
15 minutes (the first 10 were spent fighting...) she said:
"Oh ya, I got it.  Is that the trick?"   I said yes.  
Then she went back to composer and continued to work. *sigh*  
When she had a problem again (table alignment), I helped her
re-write the page (I refuse to edit stuff from those tools)
from scratch.  It didn't take long.   Once again, we spent 
about 15 min fighting about it, but then, in about 10 minutes
whipped up a pretty table.  After that she went back to 
composer.... 

Well.  I thought that I had completely failed, so I left.
Then, about two weeks later I went over to visit, (hadn't
received any more pleas for HTML help...) and I found her
using an editor to hand create the HTML! I was a bit
stunned.   She said writing HTML in an editor directly
was "easier".  She quickly added that composer is good 
too, but only to "find what I want".  She uses it to 
'draw' what she wants, looks at the 'view source' and
then ALT-TABBS over to the editor to do the 'real' work.

Anyway, seeing this, I tried another experiment.

I asked her to fill out an invoice.  She went to a web
form, filled it out, pressed enter, and it showed her
the XML.

She then went over to the EDITOR (showing off) and
put in a correct SGML rendition of it, negelecting
all of the 'obvious' end-tags. 

SUMMARY:

  I don't see XML being used directly by humans, 
  however, I do see SGML in use several years
  from now replacing data entry forms.

Thus, 

  Human -> Form -> Form Processor -> XML -> XML Validation -> Business Processes

becomes...

  Human -> SGML -> SGML Expansion -> XML -> XML Validation -> Business Processes


Then, once it's in XML, then it can be validated
for the business processes it participates.

I see SGML being used in business processing not
as a validator, but as a way to introduce the
shortcuts necessary for a productive data entry
personell to get the business information
'the way they want it'.

Then XML can be used where it shines, in back end 
validation and processing.

? Clark

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)