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


Help: OASIS Mailing Lists Help | MarkMail Help



   RE: SGML On The Web

[ Lists Home | Date Index | Thread Index ]

To understand some of it, look at the lists of the 
decision makers and discover how many of them:

1.  Were actually involved in creating a markup-based 
hypertext browser (only a handful).

2. Were experienced with a hypertext browser, markup 
or not, before HTML (a goodly number.  IADS, EBT, 
etc. had been around awhile and SoftQuad had already 
licensed a browser).  BTW: independent links (ilinks) what 
you now call out-of-line links, had 
been implemented for at least two of these systems 
and what you now call simple links but were 
then context links (clinks), for more.

3. Were actively creating formatting engines based 
on DSSSL (say James with Bosak fronting)

4. Were actively creating engines for relational databases 
(say Tim, maybe others).

5. Knew HTTP like Genesis chapter one ("In the beginning").  
A small minority.  The majority of SGMLers were not network 
experts and the real breakthroughs that led to 
the web were network designs for HTTP, not markup.  The 
Web is HTTP.  Markup is an afterthought.  PDF proved it 
not to be necessary.

Everyone that I remember working the project understood 
HTML.  HTML is a no-brainer.  That is why it works so 
well for those who need that sort of thing and why it 
will never die because there is an endless and ever 
renewing supply of web newbies.  It ain't a crime; 
it is the birthrate at work.

SGML would have worked ok but we still would have 
tossed out features and done away with subdocs.  All 
of the markup hypertext engines I was aware of had 
simplified SGML already.    DSSSL was way too 
complicated and HyTime was too obscure and had enemies 
from all camps.   That is why the syntax spec got all 
the momentum.  It was SGML--  the bit everyone had 
in common.  That was the easy bit .

Even that took a lot of wrangling because agendas 
surfaced and some were implacable (the one that kept 
me up nights was 'Kill all the DTDs').

What you have been lamenting ever since is the 
common web framework design.  It can't go forward 
fast because of HTML.  The 
model is set, the software has shipped, the 
mindsets have gelled, and the kudzu has infested.
Things will improve but ever so slowly because if 
you've ever farmed over kudzu, you know it can only 
be killed an acre or so at a time over several 
seasons if you don't want to also poison the ground.

Because of HTML, we have essentially a one client 
world and most of what XML does well, works on 
the client.  The backend is middle-ware mediated 
adaptors to what are mostly relational db servers. 
So the action is in the access layer and the client. 
The client is frozen in the "won't give up HTML 
until they pry it from our dead cold fingers", 
so that leaves the access layer.



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

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