In a comment to this question, Japanese SAPCOL digital typesetting machines are mentioned, which apparently used a PDP-8, and contained fonts for lots of Kanji.

What were those SAPCOL machines? What is their timeline? How were the fonts stored? Any other details?

2 Answers 2


A bit of googling finds this Japanese article, and DeepL does a really impressive English translation, which I am mostly just going to quote, with a bit of corrections.

Timeline overview:

  • 1960: prototype of fully automatic SAPTON typesetting machine announced
  • 1962: SAPTON-F numeric-only typesetting machine delivered to Japanse Defence Agency
  • 1965: SAPTON-N fully automatic typesetting machine for newspapers announced
  • 1966: First SAPTON-N delivered
  • SAPTON-P for books and magazines developed in parallel with above
  • 1969: First SAPTON-P delivered
  • 1969: SAPTON-A system for general printing announced
  • SAPCOL-D editing and typesetting software for SAPTON-A, using a PDP-8, developed together with SAPTON-A
  • 1970: First SAPTON-A system, consisting of SAPTON-A5260 and SAPCOL-D1,delivered.

From the SAPTON-N on, the SAPTON machines used a high-speed rotating character disk (picture in the link) and apparently some kind of optical transfer mechanism to do the actual printing.

The SAPTON-N3110 was

a fully automatic typesetter that reads CO-59 paper tape, selects the appropriate character from a high-speed rotating character disk with a capacity of 2,304 characters, and types 15 characters per line vertically on a 35mm wide non-perforative roll. The machine prints 15 characters per line in vertical format (the number of characters in a newspaper) on 35mm wide non-perforative rolls. It had one typeface, one character size for the main text, and a printing speed of 300 characters per minute (equivalent to three times the speed of a fully automatic imprinting machine).

The SAPTON-P was

equipped with a single character disk (with a capacity of 2,880 characters, the character disk can be replaced using a cassette system), image rotation prisms for vertical and horizontal typesetting, four character sizes (10-14Q), and a mono external character insertion device (up to 50 characters). One-sided perforated rolls of 2 to 10 inches wide (7 types) were loaded, and the printing speed was 300 characters per minute.

The SAPTON-A5260

held 6,400 characters, including six European fonts, a ruby font (smaller Hiragana written atop on Kanji), symbols, and numbers, on two character disks, one in Mincho and one in Gothic, and was equipped with an image rotating prism, six character sizes, and a mono external character insertion device. The SAPTON-A5440 was equipped with four character disks.

Finally, the SAPCOL-D1,

the editing and typesetting software for SAPTON-A, used DEC's PDP-8 minicomputer (16 KB main memory, 64 KB magnetic disk). At that time, minicomputers did not have an operating system (OS) equivalent, so we developed our own OS equivalent program and simple language, including multiple control of I/O devices and memory management including auxiliary storage.

SAPCOL-D1 added to the typesetting functions of SAPTEDITOR-P many processing functions necessary for general printing, such as headings, tabs, mixed typesetting of different sizes, mixed typesetting of Japanese and European text, ruby, separation prohibition, continuous numerals, and selective printing of red correction parts.

That's probably 16 KWords (not KB), or 4 core memory modules for the PDP-8, which was a common size. As for the "magnetic disk" (磁気ディスク), an RK01 stores 256 KBytes (not words, but that's still at least 128 KWords), and a harddisk stores several megabytes, so the "64 KB" is possibly an inaccuracy on side of the author. DEC tapes, which were more common than floppies, stored 184 KWords.

Anyhow, that means fonts were never stored in memory or on other storage, but always came from optical disks in the actual printing mechanism, and not in the copyediting part of the system.



In photo typesetting four basic generations can be distinguished

  1. Generation (1950s): Modified Linotype style machines. Here the metal matrices were modified to carry a character negative to be projected, but otherwise work like before.

  2. Generation (1960s): Purpose build photo typesetting. A smaller examples is the Diatype, but large integrated systems existed as well. Font images are stored on glass wheels or plates, each with a set of a given type. Either direct controlled via keyboard or by numeric input. Were early models still use mechanical gear, electronics start to appear.

  3. Generation (1970s): Computers take over the editing process, handling all metadata, but not the font image, which is still produced from a glass master. Metadata are font metrics like character width, height, kerning, etc. used to calculate exact positioning of each type. The result is used to control a printer.

  4. Generation (late 1980s): Fully digital typesetting. Now computer programs not only control the typesetting process, but also provide all font data as digital images, directly feed to automatic exposure.

The SAPTON/SAPCOL systems are a family of second and third generation photo typesetting machines.

The base are the SAPTON systems, mechanical systems, functional similar to a Diatype, that can take numeric input (paper tape, direct interface) to select characters from an optical font disk. Classic Japanese script supports less complex automation as text is written in columns and characters are of a fixed width and height. Horizontal and proportional writing are rather new developments and have been added to later iterations (SAPTON-P,-A). These are all second generation devices.

SAPTON-A in combination with a computer (PDP-8) and typesetting software (SAPCOL) constitutes a third generation system. All editing is done on the computer prior to printing. The computer system handles all font parameters to do so, but not the character image itself.

  • 1
    Gen 4 machines came in slightly earlier than that. The Linotron 202 — the antagonist in the always-amusing Bell Labs “How We Spent Our Summer Vacation” paper from early 1980 — with digital vector fonts was introduced in 1978. The next generation — CTP, direct computer-to-plate etching by laser, skipping the process camera stage — came in during the 1990s. It was a big part of my career, but not relevant here.
    – scruss
    Feb 19, 2021 at 21:32
  • @scruss Yes and no. While the 202 came in 1978 (79?), it wasn't until the second half of the 1980s that this became a widely used technology. The same is for all other dates. First photographic setting was done as early as the 1920s. In fact, the switch from 3rd to 4th generation was quite slow in comparision. Maybe because 3rd generation already realized most optimization potential thrucomputerization.
    – Raffzahn
    Feb 19, 2021 at 21:53
  • Maybe. But CTP was the biggest optimization: skipping the photographic stage was a huge saver
    – scruss
    Feb 19, 2021 at 22:56
  • @scruss No doubt, but much like 4th generation, it was an optimization that did improve/straighten the production process, but didn't offer much additional features on the user side. So the exchange against existing machinery only happened when those were ageing or new production was set up.
    – Raffzahn
    Feb 19, 2021 at 23:00

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .