Do SGML and XML co-exist?

Len Bullard cbullard at
Fri Nov 19 01:26:10 GMT 1999

Paul Prescod wrote:

Let me say, I never invented anything that has to do with SGML and 
markup systems.  I am and was a user of such systems, just a musician 
who had this strange notion that timing and synchronization were key 
to the problems of enterprise engineering.  That's all.  I am  
an older user than most of you and because of that, I was a minor 
observer of some of the history and people who did invent markup.  It
a privilege and an honor.  It was at times, thrilling.
> I don't believe that. There was a requirement for an e-commerce and data
> interchange meta-language. We were there first. 

Interestingly, it had been brought up before XML was gleam.  Every time
suggested to the EDI committees that markup was preferable, it was 
dimissed as an irrelevant text format.  When I was at the PDES meetings 
when SGML was first considered, it was only as a string alternative 
to the somewhat byzantine document model they were building.  No, Paul, 
two things made the EDI folks take it seriously:  MS was backing it 
and it was going On The Web.  If I can find Rita Knox, we can probably 
still get an authoritative, deeply technical, and completely irrelevant 
analysis of why markup is a bad idea.  If Betty Harvey remembers, at one 
point, HTML was all that mattered.  If I am honest, I thought Hytime 
would change the world.

We all have those moments.

> The client side (if it
> ever materializes) will be gravy. Microsoft probably accelerated the
> process but there was a desperate market need for a faster way to
> describe data interchange languages than BNF and SGML really should have
> been there several years ago but for the stigma and the DTD dependency
> issue.

The stigma is the interesting issue.  We already knew how to work with 
well-formed files and without DTDs.  We struggled against the SGMLWay, 
the consultancies, and the fact that before HTML blew away all other 
competitors, we could not get a hearing.  Why?  Politics and money.  

I had the good fortune in Orlando to watch Dr. Goldfarb take on Larry 
Welch from NIST.  It was incredible to see this balding bearded 
Jewish lawyer with cowboy boots and a string tie taking notes 
on that long yellow legal pad, drop his glasses to the end of his 
nose, then devastate Welch.  Welch hateed SGML.  He used his
girth and position to fight it at every turn.  That was what SGML was 
up against:  committees, power brokers in Washington and elsewhere,
with petty empires and overarching ambitions.   

Most of you on this list probably don't have the real picture of 
what you owe Goldfarb.  Considerably more than a beer, 
but if you meet him, consider yourself lucky and buy him one if Linda 
isn't looking.  He built this language, folks.  Some of you are 
getting credit and pocket change, but he made this happen.  Not Jon, 
not Tim, not TimBL.  I respect those guys, but Goldfarb paid for this
Respect is owed; respect is due.

So why did some of us go with the SGML On The Web project:

1.  It was the right thing to do.  HTML is an SGML application.  I had 
to defend it in DC against every comer.  Even when I knew there were 
better designs, it was evolving fast and at a certain point, it met 
the requirements, so Selah.  Shake hands and move on.  I stood in front 
of the assembled IETM cognoscenti of Lockheed Martin in Orlando and told 
them, yes, we spent a lot of money, we've done a lot of good work. 
it training.  Now, get with the program.  The Web wins.  It hurt, but 
it was the right thing to do.  As a professional, one will have those 

2.  We needed a means to focus the SGML community.  The long time
for DSSSL to be completed (the real tent pole), the suffering with FOSI 
based systems so complicated we could never get them in high producion, 
the mysterious, inscrutable and ultimately overbuilt HyTime standard,
all of 
this had gone on too long.  Without the web and the money, there was no
the excellent conceptual foundations of SGML would have survived.

3.  It wasn't money, at least, not for me.  Haven't made a dime on it.  
Probably won't.  I've invested a lot of time in VRML and the fruits 
of that are just emerging not as money, but in that next year, there 
will be a markup version of animtated 3D. I helped make that happen.
Why?  It is the right thing to do.  Wall-to-wall markup systems have 
the potential to make the Total Cost of Ownership (TCO) of complex 
information come down.  Real time 3D is a very key piece and we have 
known that since at least the eighties.  To happen we had to get 
off expensive platforms.  Today we have 50 cent memory and two 
dollar processors.  Time for Phase II.

It takes a long time to do some kinds of work.  You will meet a lot of 
people, eat a lot of bad food, and stay in some nasty rooms.  You will 
have a lot of laughs,  some mindbending arguments, shed a few 
tears, and only win some of the games.  You will give up some things 
you hold dear, risk friends, alienate allies, even walk away when 
you are the impediment, not the solution.  Ask yourself why and be
about it.  Money only goes so far. 

It ls what it has always been:  we have to protect information, 
as Charles said, "from folks like us".   Why?  The humans matter 
more than the machines.  They need information, and for that,
must be free of implementation.  It isn't content vs presentation.  
That is an old red herring.  It is content independent of
The reason some us understand DTDs.   SGML and XML 
aren't inevitabilities.  They never were.  They are there because 
two and now three generations of information pros and users fought 
for them.  There Ain't No Justice Unless You Make It So.  We aren't 
martyrs;  we are pros.  We take pride in it.  It is, as Lawrence 
told the Sheik, "our pleasure" to do it.  We do it for people. 

But as Megginson notes, without the money, it isn't comfortable and 
with the money, it is a lot more exciting if a lot more crowded. I just 
won't sit here and let history be rewritten by newbies and those 
with a platinum axe to grind.  I accept XML.  How many of you on the 
list accept that XML is SGML, and because of it, has a proud and well 
respected heritage made by people whose names are fading faster than 
the sun over the Pacific?  Respect is owed; respect is due.
> Admittedly, had Microsoft come out with an XML *competitor*, XML might
> have been in trouble. But if they sat on their hands, we would still
> need to solve the Internet e-commerce issues with or without them and I
> don't see any other proposals for data interchange meta- and schema-
> languages.

Common LISP comes to mind.  S-expressions, as Dan Connoly told me, are 
the highest form.  Why not LISP?  Well, because HTML wasn't written 
in LISP.  As the twig is bent, so grows the tree.


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

More information about the Xml-dev mailing list