Mason, James David. “SGML at 40.” Presented at Balisage: The Markup Conference 2026, Washington, DC, August 3 - 7, 2026. In Proceedings of Balisage: The Markup Conference 2026. Balisage Series on Markup Technologies, vol. 31 (2026). https://doi.org/10.4242/BalisageVol31.Mason01.
Balisage: The Markup Conference 2026 August 3 - 7, 2026
Balisage Paper: SGML at 40
James David Mason
James David Mason, originally trained as a mediaevalist and linguist, became a writer,
systems developer, and manufacturing engineer at U.S. Department of Energy facilities
in Oak Ridge in the late 1970s. In 1981, he joined the ISO’s work on standards for
document management and interchange. He chaired ISO/IEC JTC1/SC34, which was responsible
for SGML, DSSSL, Topic Maps, and related standards, for more than 20 years. Dr. Mason
has been a frequent writer and speaker on standards and their applications. For his
work on SGML, Dr. Mason received the Gutenberg Award from the Printing Industries
of America and the Tekkie Award from GCA. He recently retired from working on information
systems to support manufacturing and documentation at DOE’s Y-12 National Security
Complex (Y-12) in Oak Ridge, Tennessee.
ISO 8879, the Standard Generalized Markup Language, SGML, was published 40 years ago.
A brief history of SGML and related specifications seems appropriate at
this conference.
Forty years ago this summer was an event without which we probably would not be holding
this conference today: the publication of ISO 8879, the Standard Generalized Markup
Language, SGML.
The idea of generic coding in publishing was not new in 1986; it had arisen almost
as soon as programmable computerized document formatters were introduced in the 1970s
– indeed enabling generic coding was a major reason for making such software programmable.
Macro systems were developed for Scribe, Troff, IBM’s Document Composition Facility,
and most notably for Donald Knuth’s TeX, the latter case being Leslie Lamport’s LaTeX,
which appeared at almost the same time as SGML.
Generic coding took another important direction with Bill Tunnicliffe’s organizing
of the GenCode committee at the Graphic Communications Association, a gathering of
end users seeking a generic solution independent of any particular formatting software.
Their work led to the forming of an ISO study by the Experts Group on Computer Languages
for Processing Text, and that evolved into the committee that created SGML.
Before the completion of the standard, GCA published the balloted 1983 draft as GCA
Standard 101, GenCode; and Tunnicliffe convinced the Department of Defense to adopt
it as a trial usage.
Figure 1: Cover of the Gencode Standard
This set off a surge of development that meant that there were working implementations
of SGML by the time the standard was finished.
SGML differed from earlier generic projects in that it did not simply catalogue generic
elements of text, but rather gave users the tools to create their own sets of elements.
Much of the credit for the design of SGML belongs to the project editor, Charles Goldfarb.
He should be remembered particularly for the choice to create a language and for giving
it a formal definition through a production grammar.
Figure 2: Cover of the ISO 8879 1996 Standard
SGML also differed from other information technology standards in that it was created
by end users, for end users, not by vendors or software designers, although many of
us on the committee had programming experience. We did this because we needed it,
not because someone wanted to sell us something.
Figure 3: Example of SGML Features
(This table, which officially is an example of some SGML features in the appendices
to the ISO standard, is more or less the membership of the ANSI committee, with a
few ISO and honorary members thrown in. It’s Goldfarb’s way of memorializing some
contributors in a supposedly anonymous ISO standard. The list dates to 1983, before
I came to chair the ISO committee.)
While this background perhaps explains some of the quirks of the standard, it still
drives the current world of markup, as can be seen by the typical population of this
conference.
The publication of ISO 8879 set off a cascade of standards development, user implementations,
and GCA conferences. Balisage has been the heir to the latter.
Having finished SGML, the ISO committee began work on the Document Style Semantics
and Specification Language (DSSSL). Though it is little used today, DSSSL led to the
development of XSLT and XSLFO, which many of us use.
When challenged to design an SGML representation for the Office Document Architecture
standard, Goldfarb came up with the concept of Architectural Forms, and that led to
the development of HyTime, SMDL, and Topic Maps, about which we will hear later this
week. And philosophically, if not technically, this pattern has influenced DITA.
And of course there is HTML. When he was trying to create an encoding for the Web,
Tim Berners-Lee turned to Anders Berglund, CERN’s representative to the ISO committee,
who recommended SGML syntax and the basic element set.
Then there is this conference itself whose name, Balisage, is the last word of the
French title of the SGML standard.
