W3C home > Mailing lists > Public > www-xsl-fo@w3.org > October 2002

Re: xsl-fo first anniversary

From: Sharon Adler <sca@us.ibm.com>
Date: Mon, 21 Oct 2002 12:11:40 -0400
To: eliot@isogen.com
Cc: www-xsl-fo@w3.org, www-xsl-fo-request@w3.org
Message-ID: <OF2B9820EA.56704117-ON85256C59.0056ECF6@us.ibm.com>






Not too bad for a kid Eliot.  Hope you all will accept my "history lesson"
- I am feeling nostalgic these days.

1. DCF and TeX development was independent to my knowledge.  DCF was late
70's around 78 (release 1) - same macro concepts implemented in Waterloo
script by Bruce Utley and Roger Watt - called SysPub.

2. 1960's GenCode was really late 1970's GenCode and it was a Macro
replacement (p1, p2, etc.) for standard paragraph levels.  Bill Tunnicliffe
and Charles Lightfoot spearheaded this effort during the 70's.  I became
GenCode Committee chair for the GCA in 1979 when the GenCode committee
merged with the ANSI committee doing SGML.  This became the ISO committee
doing SGML in early 80's.

3.  Brian Reid did Scribe a typesetting language that also had a macro
front-end and he spoke at a GenCode Conference from the GCA in 1979.

4. DSSSL started in 1986-87 (overlapping with the final "tweaking" of
SGML).  FOSI started a bit later since it  was clear that the DSSSL effort
would take a while and the DOD was in a hurry.  FOSI at first seen to be a
placeholder for DSSSL.  (We spend a great deal of time fighting the ODA
wars).

5. Eliot I joined IBM in 1985 and from the beginning we were looking to
build an SGML-based composition engine - work on last effort was late 80's
early 90's which ceased in '92.

6. James joined the DSSSL committee - I literally dragged him to a meeting
at Hatherley Manor after I met him when I gave a talk on DSSSL to a local
SGML UK user's group.  Wouldn't let him go back to London.  He stayed -
took over much of the writing and the rest is history.

7.  DSSSL & HyTime had to have groves since we both had entirely different
data models for SGML - pre-dates the Infoset - but had some required
structures so you had something to hand your hat on.  Took an extra year to
get DSSSL to ISO standard status to get groves into DSSSL and then the
DSSSL transformation language was never implemented except by some rare
souls who implemented small subsets of it.

So you see Eliot you were not really inaccurate - had a few dates wrong,
but only us old fogeys remember dates.

Cheers everyone.

Sharon



Sharon C. Adler
 Senior Manager, Extensible Technologies
 IBM Research
 PO Box 704, Yorktown Heights, NY  10598
 tel:  914-784-6411 t/l 863
 fax: 914-784-6324



                                                                                                                               
                      "W. Eliot Kimber"                                                                                        
                      <eliot@isogen.com        To:       www-xsl-fo@w3.org                                                     
                      >                        cc:                                                                             
                      Sent by:                 Subject:  Re: xsl-fo first anniversary                                          
                      www-xsl-fo-reques                                                                                        
                      t@w3.org                                                                                                 
                                                                                                                               
                                                                                                                               
                      10/19/02 09:46 AM                                                                                        
                      Please respond to                                                                                        
                      eliot                                                                                                    
                                                                                                                               
                                                                                                                               




Dave Pawson wrote:
> At 21:57 18/10/2002, W. Eliot Kimber wrote:

>>1960's: gencode projects start as a way to standardize typesetting codes.
Basic ideas of generalized markup for documents start to form
>>1970's:
>> - Goldfarb, Moscher(? sorry, don't have SGML Handbook to hand), and
Laurie develop GML at IBM in order to satisfy requirements of legal
document search and retrieval. GML implemented on top of IBM's Document
Composition Facility (DCF).

> Edward Mosher and Raymond Lorie, yep found them.

Thanks--I should have been able to remember their names better, but I am
hideously bad with names.

>> - Various large publishers, such as Boeing (where Sharon worked) and the
U.S. government start using GML to solve large-scale publishing problems.
Quickly realize the limitation of having markup language bound too tightly
to underlying typesetting language.
>
>
> So this  was the time the split began between style and content?

Yes.

>> - TeX is developed. LaTex developed--similar to GML in that it is a more
semantic macro language on top of a low-level formatting language (not sure
of the exact timing here).
>
> Any direct relationship between gml and tex or were they parallel
> developments?

I don't think there was much cross-pollination here, but I don't really
know. They were completely parallel as far as I know. I know that by the
time we were doing the SGML thing at IBM we were looking at TeX as a way
to implement composition, but that was late '80s, early '90s.

>> - HyTime standard work started. Charles Goldfarb editor and driving
force
>> - FOSI spec developed to meet U.S. DoD requirements (not sure of exact
timing here). Implemented by Datalogics and Arbortext.
> Was Paul with Arbortext then?

Yes. Paul was one of the founders of Arbortext.

>>Early 1990's:
>>
>> - IBM starts developing SGML replacement for it's GML application
BookMaster (used for 90% of IBM's publications).
>
> Oh! News to me.

The IBM Information Development document type survived my departure,
albeit in much revised form (can you say "naive over-engineering by an
over-enthusiastic kid"?) and is still being used today as far as I know.
Ask Ken Holman about the first time he and I met....

>> - HyTime standard published.

> Sounds like a lonely interjection Eliot.
>   what was the motivation for this work please?

There is no accurate short answer to this question, but to oversimplify,
the goal was to enable the SGML representation of hypermedia
information, from abstract source to recordings of performance
instances. My personal motivation was focused primarily on solving
problems of large-scale hypertext as required by the types of complex
systems of technical documents I'd been working with at IBM. Steve
Newcomb's motivation was focused primarily on solving problems of
representing music and performance in general (e.g., opera). Charles
wanted all of these things. The HyTime standard itself was an outgrowth
of the Standard Music Description Language (SMDL), an attempt to define
a generic markup language for music. The way Steve N. tells the story,
the U.S. Dept. of Defense came to a meeting of the standard committee
and said essentially "if you can represent operas you can represent
battles and we need a way to represent battles but we can't go the
Pentagon and suggest they apply a music standard, so if you could break
out the generic bits and make that a separate standard, that would be
great." And so they did. And SMDL continues to languish to this day as
other activities have pushed it to the background for those involved
(sadly, there does not appear to be any money for the generic
representation of music, but there is lots of money for topic maps, FO,
and workaday applications of XML--so it goes).

>>1995:
>>
>>- HyTime and DSSSL camps realize they must come to agreement on the
fundamental data model for SGML documents. Groves invented to solve this
problem.
>
>
> And nobody got groves either?

Correct. Nobody got groves. But maybe that is changing, if the last
Extreme conference is any indication. Hope springs eternal.

>>1996:
>>- DSSSL standard published. James Clark releases Jade DSSSL
implementation.
>>
>>- John Bosak starts "SGML on the Web" activity with the express goal of
enabling high-quality rendering of SGML documents in Web browsers.

Just realized I mispelled Jon's name. Sorry about that.

>
>
> That's interesting. I started to monitor xml-dev in early 97, and hadn't
> realised that was a goal, certainly not the 'high quality print' aspect.

It was. Jon was very clear about that to us (like my self, Jon is
fundametally a technical writer, a book person). But it also became
immediately clear that that was not where XML was going and so he didn't
push the issue. But fortunately, those who care about high-quality
rendition from XML didn't give up the struggle. It's kind of cool how
all the non-print uses of XML have provided both the infrastructure and
the time for the print uses of XML to both mature and be affordable. We
certainly never would have predicted that at the time.


> Thanks Eliot. Much appreciated.

My pleasure. I hope it's not too inaccurate.

Cheers,

E.
--
W. Eliot Kimber, eliot@isogen.com
Consultant, ISOGEN International

1016 La Posada Dr., Suite 240
Austin, TX  78752 Phone: 512.656.4139
Received on Monday, 21 October 2002 12:16:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.1 : Tuesday, 21 September 2004 12:20:09 GMT