The IBM 2250 Display Unit was originally shipped with the IBM 1130 computer, introduced in 1965. The 2250 could also be attached to IBM 360-series mainframes, as ours was to the 360/91. Like most IBM terminals, attachment was via control unit (or in this case, direct channel) rather than communication port.
The 2250 was the "first commercially available graphics terminal" if you don't count the DEC PDP-1 display (1961). As late as 1971 there were only about 1000 interactive CRT graphic terminals installed in the USA, compared with 100,000 line printers, 50-100,000 Teletypes, and 70,000 "alphanumeric terminals" (such as the 2260) (CACM Vol.14 No.1, January 1971, p.60).
Columbia's 2250 was intended for use by physicists to read, display, and interact with cloud chamber photos, from ongoing physics experiments in the machine room (the 360/91 was partially funded by physics research grants). From Peter Capek of IBM, who was here in the 1960s: "The 2250 was a totally different beast [from the 2260]. I don't know about the cost, but $100K doesn't seem implausible for the channel-attached version. It was a separate product, sold either for use with an 1130 (stand-alone small machine, mostly targeted at engineering/design applications), as a design workstation for use with a S/360 (channel attached), or in a special version that was built into the /91 console and not an option. I think the latter version cheated a bit on the refresh memory size and flexibility...
"It mostly was not used [at Columbia], as I recall, because the support for it in the operating system was almost non-existent. (No other 360 beside the 91 and 95 (195?) had this, so IBM didn't put a lot of effort into it.) It was a vector display with drawn characters and a light pen, so you could point at something on the screen, drag a crosshair, etc. I can't recall ever seeing anything other than characters on the machine at Columbia, and that only very rarely. I believe the 2250 per se did predate System/360, but I'm quite sure it could not be connected to a709x system. I also can't recall it ever being used for any HPD function at Columbia, perhaps in part because of it's location as the operator's console; can't have physicists wandering arount there. The console function was provided by the 1052, a mechanical Selectric-typewriter like printer/keyboard." [In the 1970s, after Peter's time, it was indeed an operator display, used for showing active jobs.]
Hank Butler was a longtime member of the Machine Room operations staff in various roles, such as chief Tape Librarian. He is shown playing a game of chess against the 360/91 on the 2250 about 1971: an early computer chess program called CCCP (Columbia Computer Chess Program) written by Steve Bellovin (now Computer Science faculty member) and Columbia Computer Center programmers Aron Eisenpress, Ben Yalow, and Andy Koenig (son of Seymour Koenig, director of IBM Watson Lab at Columbia University 1952-1970).
Peter Kaiser, who was also here in the 1960s, confirms that the 2250 was new with the 360/91, whose predecessor, the 7094/7040 Directly Coupled System, had "an attached device, not a 2250, used to read and display cloud chamber photos. It was still there when I arrived in 1967 and although there was some discussion of hooking it up to the 360 systems, I think that never happened."
Here's a picture of the 2250 from a 1968 Sikorsky Aircraft ad, showing an enlarged version of the stroke-drawn character set, the light pen, and the function keypad. The article: Engvold, K.J., and J.L. Huges, A General-Purpose Display Processing and Tutorial Systems, CACM Vol.11 No.10, October 1968, pp.697-702, describes the ADEPT (A Display-Expedited Processing and Tutorial) system, an interactive simulation, programming, information retrieval, and teaching system for OS/360 using a 2250 display station. An early and very expensive example of a graphical user interface, with user input from keyboard, function keypad, and light pen, and output in text and graphics, implemented entirely in software.
IBM followed the 2250 with a 3250 vector graphics unit, and later (mid 1980s) by the 5080 raster graphics station.
Also see: The IBM 2260 Display Station.
Thanks to: Steve Bellovin, Department of Computer Science, Columbia U, for the two Columbia Machine Room photos, 10 March 2021.
|Columbia University Computing History||Frank da Cruz / firstname.lastname@example.org||This page created: August 2003||Last update: 11 March 2021|