A rather different recollection of the origin of DTP

As someone who learned to set lead type by hand, using a composing stick, and has remained active in printing and graphic arts in Silicon Valley ever since, I find this article distressingly inaccurate. If, by "desktop publishing" (DTP), the author is referring to technologies which replaced phototype and "mechanicals" (pasteups) in commercial printing, then it was the combination of the Macintosh, the Adobe PostScript language, a handful of Adobe Type 1 PostScript fonts, the PostScript interpreter added on to an existing Linotype imagesetter which was then named the "L-100," the PostScript-compatible Apple Laserwriter (and its PostScript driver, which was also used for the L-100), and Aldus PageMaker which began the DTP revolution. MacPaint, MacDraw, and MacWrite also played important roles.

Within a year of the introduction of these items, the market for phototypesetters collapsed.

QuarkXPress soon replaced PageMaker as the page layout program of choice. Adobe Illustrator and Aldus Freehand (formerly Altsys Virtuoso) added vector graphics creation to the mix. Much later, version 3.5 of Photoshop made professional bit-mapped graphics editing possible. Altsys's Fontographer spurred the creation of new fonts. Intense customer pressure and the prospect of legal action led Adobe to allow other companies to produce Type 1 fonts. (Adobe's president, John Warnock, cried at the press conference announcing this decision.)

DOS and Windows users got Ventura Publisher (which actually ran in Digital Research's GEM environment). Ventura was loathed by everyone except some marketing types who were forced to use PCs.

SGML played almost no role in DTP, except in one niche--software manuals and similar docs. Interleaf and FrameMaker spoke SGML, but commercial printers--even in Silicon Valley--rarely had to deal with this. SGML played no role in typesetting prior to this, because every phototypesetter manufacter had one or more proprietary markup languages. The typesetting "frontends" which produced either punched paper tape or were wired directly to phototypesetters, were either hardwired for one such language or were programmable to allow their use with several languages.

Punched paper tapes, depending on manufacturer, had anywhere from 5 to 8 holes per line, with a corresponding difference in language complexity. Some phototypesetters used other media. Linotype modified its machine to use film negatives instead of brass molds. Alphatype's Alphasetter, the highest-quality phototypesetter ever, used reel-to-reel audio tapes.

WYSIWYG ("what you see is what you get") was available long before DTP. Even the low-end typesetter manufacturer, Compugraphic, offered a WYSIWYG terminal for its Editwriter machines. Similarly, typesetting houses were outputting phototype in multiple columns--even on film--long before DTP.

As soon as microcomputers became available, they were used to drive phototypesetters. I was involved with a San Jose graphics house which used an Altair 8800 to drive a Photon (high-end) phototypesetter in the mid 70's. In the 60's, the folks playing in the big leagues, such as RR Donnelley and Black Dot, were using minicomputers to set type, soft proof it (often WYSIWYG), and drive optical and CRT typesetters.

The bit about the graphics tablet was interesting and new to me, although the idea of using bitmapped type output on a dot matrix or laser printer for printing died almost as soon as PostScript became available.

and before DTP

let's not forget that it was the organisation of metal typefaces, with the small letters low down near the workspace, and less-frequently-used CAPITAL letters placed in the harder-to-reach uppermost box (or case), that gave us the terms uppercase & lowercase that we still use today.