dns@sq.uucp (David Slocombe) (08/02/88)
In article <62137@sun.uucp> tut%cairo@Sun.COM (Bill Tuthill) writes: > >Could SGML possibly be misnamed? Was it hopelessly naive of me to assume >that the word Markup in its name indicates it is used for markup? > When I first went to work for a newspaper in 1966, we "marked up" copy with phrases like "Times Bld Ext 24/28 centre", in pencil, with a circle drawn around all such markup to distinguish it carefully from the text of the story. (This story was going to be keyboarded by typists working on paper tape punches with no visual feedback -- "blind perforators". They were paid by the keystroke, with their rate based on their average speed, so you had to make everything very, very clear.) That sort of thing, and other marks that are not so easy to describe on an ascii keyboard, was called "markup". The perforating typists, when they saw these encircled things, keyed in special codes comparable to troff requests. Sometimes they also keyed special codes which were called "formats" but were really just simple macros (non-recursive). The programmers involved in typesetting tended to refer to all these codes as "markup", by extension. In the 1970s, the group of mainly large typesetting firms and in-house typesetting departments that formed the membership of the Graphic Communications Association (by "large" I mean big enough to have a Xerox 9700 printing proofs and have a SECOND 9700 on hot-standby!) developed the notion of "Generic Markup" or "Generic Coding", meaning that a standard set of codes would be used in keyboarding, independent of the particular typesetting equipment that was going to be used. This meant no retraining of operators when they changed equipment, and it also meant that they could start thinking about having their clients supply them with machine-readable manuscripts. People like the purchasing agents of the U.S. DOD and the Canadian Queen's Printer loved this idea because it meant they could get manuscripts keyboarded with the generic markup in place -- cheaply -- and then shop around for the best competitive bid for typesetting the machine-readable file. Without Generic Markup, the typesetting firms would insist that they had to re-keyboard the document to match their typesetting equipment (at great expense, of course). So both U.S. and Canadian governments became instant converts to Generic Markup, and SGML soon thereafter. The codes described in the Generic Markup literature were still largely typesetting-specific (for one thing they tended to be state-changes in the middle of a stream of text: process oriented), but "generic" in the sense that the typesetting staff had some control over things at a later stage in the production. Hence you had "<h1>" for "heading style 1", and <it> for "italic [but the type-family to be specified elsewhere]". In effect they were calls to macros and that is how their processing was generally implemented. This was still called "markup", but you can see a trend away from stylistic details. "Heading style 1" may in practice expand into "\f(TB\s(14" but it can also be considered a description of a line of text as a logical element [of the document] of the type "heading-1". From "Generic Markup" to SGML was a relatively small step conceptually. (The step was taken around 1980 when the GCA GenCode committee merged with the ANSI SGML committee.) The codes you use to "mark up" your document in SGML no longer have ANY visual-style meaning left, since that issue has all been punted down-stream within the typesetting shop, as it were. But the codes are still necessary if the formatting is to proceed without a hitch, so they inherit the name of "markup". ---------------------------------------------------------------- David Slocombe (416) 963-8337 SoftQuad Inc. uucp: {utzoo,utai}!sq!dns 720 Spadina Ave. Internet: dns@sq.com Toronto, Ontario, Canada M5S 2T9