Ancestors

Written by Ken Shirriff on 2025-02-01 at 18:56

What is the origin of the word "mainframe"? Digging through archives, I traced it back to 1953. The IBM 701 computer was built from "frames": power frames, a storage frame, a drum frame, and the main frame. This 1953 drawing from the Installation Manual shows the dimensions of the "main frame". 1/n

=> View attached media

=> More informations about this toot | More toots from kenshirriff@oldbytes.space

Written by Ken Shirriff on 2025-02-01 at 18:56

IBM's construction of computers from frames was a major innovation: the frames fit through doors and were transportable. Earlier computers were often constructed in place and difficult to move. Photo shows an IBM 701 at GE with frames for power and drum storage, and main frame in the center.

=> View attached media

=> More informations about this toot | More toots from kenshirriff@oldbytes.space

Written by Ken Shirriff on 2025-02-01 at 18:58

IBM's engineering documentation extensively used the term "main frame". The documentation described the location of each tube in the system. The abbreviation "MF" indicated the tube was in the main frame, as in the schematic diagram. 3/n

=> View attached media | View attached media

=> More informations about this toot | More toots from kenshirriff@oldbytes.space

Written by Ken Shirriff on 2025-02-01 at 18:59

From the 1950s to the 1980s, the word "mainframe" indicated the physical box or cabinet that held the computer. Even a personal computer had a "mainframe". 4/n

=> View attached media | View attached media

=> More informations about this toot | More toots from kenshirriff@oldbytes.space

Written by Ken Shirriff on 2025-02-01 at 18:59

At the same time, "mainframe" took on a different but related meaning: a computer's CPU (central processing unit). By the 1980s, however, this definition faded away: a book stated: "According to the official definition, 'mainframe' and 'CPU' are synonyms. Nobody uses the word mainframe that way."

=> View attached media

=> More informations about this toot | More toots from kenshirriff@oldbytes.space

Written by Ken Shirriff on 2025-02-01 at 19:00

With the development of the minicomputer and the microcomputer in the 1970s, "mainframe" moved towards its modern meaning, with computers partitioned into micros, minis, and mainframes. IBM started using "mainframe" as a marketing term in the mid-1980s; e.g. calling the S/390 "mainframe class". 6/n

=> View attached media

=> More informations about this toot | More toots from kenshirriff@oldbytes.space

Written by Ken Shirriff on 2025-02-01 at 19:00

Analyzing newspapers shows the rise and fall of the term "mainframe". The term became popular during the 1980s. At the end of the 1990s, as the popularity of mainframe computers dropped, usage of the term steeply dropped too. 7/n

=> View attached media

=> More informations about this toot | More toots from kenshirriff@oldbytes.space

Toot

Written by Ken Shirriff on 2025-02-01 at 19:01

For more on the history of the word "mainframe", see my latest article: https://www.righto.com/2025/02/origin-of-mainframe-term.html

Thanks to the Computer History Museum for letting me examine their archives.

=> More informations about this toot | More toots from kenshirriff@oldbytes.space

Descendants

Written by Mark Crocker on 2025-02-01 at 19:42

@kenshirriff they also use the word screen very differently than the modern context 🤣

=> More informations about this toot | More toots from mcrocker@indieweb.social

Written by lopta on 2025-02-02 at 02:44

@mcrocker @kenshirriff Remember when a monitor was a certain small software utility?

=> More informations about this toot | More toots from lopta@mastodon.social

Written by Vertigo #$FF on 2025-02-01 at 19:46

@kenshirriff Tagging #mainframe (see up-thread)

=> More informations about this toot | More toots from vertigo@hackers.town

Written by ezra on 2025-02-01 at 20:34

@kenshirriff have you and david talked about editing some of your posts into a book? MIT press should gobble this up

=> More informations about this toot | More toots from ezra@assemblag.es

Written by root42 on 2025-02-01 at 20:39

@kenshirriff wasn‘t CPU also sometimes not just a chip but maybe a whole PCB or even chassis?

=> More informations about this toot | More toots from root42@chaos.social

Written by Ken Shirriff on 2025-02-01 at 21:25

@root42 Yes, the "central processing unit" was defined as "A unit of a computer that includes the circuits controlling the interpretation and execution of instructions. Synonymous with main frame". Sometimes storage was considered part of the central processing unit.

=> More informations about this toot | More toots from kenshirriff@oldbytes.space

Written by Stephen Illingworth on 2025-02-01 at 21:28

@kenshirriff @root42 I've heard people referring to the case of a PC as the "CPU". I didn't like it.

=> More informations about this toot | More toots from JetSetIlly@mastodon.gamedev.place

Written by Matthew Miller on 2025-02-02 at 00:08

@JetSetIlly @kenshirriff @root42

Yes, me too. Like, as opposed to the monitor or keyboard or other peripherals. This history shows these people were more attuned with the origin of the term than we were!

=> More informations about this toot | More toots from mattdm@hachyderm.io

Written by 💞 eva 💞 on 2025-02-02 at 14:34

@mattdm oh dear, making me feel old this early on a Sunday! I remember all of those terms, from spending my after-school hours at the US Naval Academy's computing rooms. Some of my best childhood memories.

This thread is great! Thanks Ken!

@JetSetIlly @kenshirriff @root42

=> More informations about this toot | More toots from winterschon@bsd.cafe

Written by Henrik Kramselund - kramse 🍉 on 2025-02-03 at 08:34

@JetSetIlly @kenshirriff @root42 they also have used "the hard drive" for the whole PC, equally annoying

=> More informations about this toot | More toots from kramse@helvede.net

Written by Ben Galehouse on 2025-02-02 at 10:51

@root42 @kenshirriff My father had a PDP-11/45 in our basement when I was a kid. The CPU was a stack of perhaps 5 boards of discrete TTL chips (a few flip-flops or what-not per chip) and a lot of wirewrapping on the other side of the backplane to connect them. Next to that was the FPU, which was larger - maybe 12 boards if I remember right.

Keeping it working often meant identifying the chip which went bad. To assist with this, he managed to acquire an extender to hold a board out from the backplane a bit, which gave more room (though still not a lot of room) to clip on leads to connect to the oscilloscope/logic analyzer. However, using them required downclocking the CPU, which occasionally made the fault go away.

=> More informations about this toot | More toots from bgalehouse@mathstodon.xyz

Written by Thomas Amberg on 2025-02-01 at 20:47

@kenshirriff reminds me of https://www.flickr.com/photos/tamberg/54277912940 (Hohner-CPU, 50ies)

=> More informations about this toot | More toots from tamberg@quite.social

Proxy Information
Original URL
gemini://mastogem.picasoft.net/thread/113930173443705439
Status Code
Success (20)
Meta
text/gemini
Capsule Response Time
345.760655 milliseconds
Gemini-to-HTML Time
5.962927 milliseconds

This content has been proxied by September (3851b).