Forum > Other
System/360 history
440bx:
Hello,
The following is a link to a youtube video about IBM's System/360.
https://www.youtube.com/watch?v=npgvV_-Nh60
Just FYI, it's mostly pictures and narration (no fancy eye candy anywhere.)
It's not obvious today but, System/360 is in many ways the "big bang" of the modern computer era.
It's genuinely amazing that, at one time, a fully loaded mainframe was a machine with 4MB of memory (and, back then, that was a spectacular amount of memory very few organizations could afford,) On configurations that had quite a bit less memory, you had COBOL, FORTRAN and PL/1 compilers available.
It's genuinely amazing what programmers managed to do on such limited hardware.
Anyway, for me personally, that video brought a lot of good memories back. :) I hope it will do the same for the forum old timers and provide somewhat of a picture of what computing used to be at one time to the younger generation.
Enjoy!.
VisualLab:
--- Quote from: 440bx on April 28, 2024, 12:03:17 am ---It's genuinely amazing what programmers managed to do on such limited hardware.
Anyway, for me personally, that video brought a lot of good memories back. :) I hope it will do the same for the forum old timers and provide somewhat of a picture of what computing used to be at one time to the younger generation.
--- End quote ---
"Squeezing the last byte" from hardware.
Now, when I sometimes read some articles about JS or Python "engineers", my skin crawls. In particular in relation to equipment such as microcontrollers, smartphones or SBCs. Technological "debauchery and hedonism" in IT. Are we currently doomed to poorly made, temporary, shoddy, makeshift solutions?
TRon:
Thank you 440bx,
That was indeed a very nice trip down to memory lane. Without it's existence (and IBM as a whole) we wouldn't be there where we would be today.
As indicated by VisualLab it perfectly shows how the rat-race for better, faster stronger is a never ending one. Ofc any attempt to stop that is at the expense of convenience but for sure restricted hardware makes you think twice about how to approach an issue and come up with an adequate solution.
440bx:
I'm very pleased not being the only one who enjoyed it.
There are so many significant advances in the making of System/360 that it almost feels sacrilegious to place one of them at the top of the heap but, IMO, I believe the development of an operating system, something which did not exist before System/360, is likely the most significant game changer that project produced.
It's not obvious but, the advent of O/Ss influenced CPU capabilities, instruction sets and memory management. Many hardware features came about to support the production of more capable O/Ss.
I found it entertaining to see the development of microprocessors mirror that process. A significant number of microprocessor features exist to support better O/Ss. It shows that the hardware guys do "live and learn" :)
MarkMLl:
I broadly agree. While IMO the "every member of the older IBM 7000 range was different" angle is overplayed, the 360 emphasised a number of points- fairly regular register set and so on- which eventually came to dominate the industry.
One quirk however is that it didn't have a hardware (return etc.) stack, which is something that almost every architecture now has and which makes implementing ALGOL-derivative languages (i.e. with recursion etc.) easier. Another was that it had limited internal protection: it was only robust if programmers could be trusted to only use precisely the facilities that IBM documented and while I believe that was progressively improved it remained a flaw into the Internet era.
I found the book at https://core.ac.uk/download/197546563.pdf useful as an historical resource (i.e. to disentangle system names etc.) and to explain the variant terminology that the IBM priesthood has always used.
There was an attempt at an FPC port some years ago, but it stalled because of friction over e.g. EBCDIC vs ASCII and the author's difficulty understanding the codebase (he was unhappy with the core developers' position that a full function cross-reference was impossible). In addition, my experience on the edge of the "historical reenactment" scene suggests that fundamental differences over file storage would have become a big thing if the target were MVS (formerly OS/360) rather than VM/CMS which is somewhat unix-like.
There is of course https://bernd-oppolzer.de/job9.htm hence https://github.com/StanfordPascal/Pascal, but I have not really explored how compatible it is with FPC etc.
MarkMLl
Navigation
[0] Message Index
[#] Next page