Slashdot Mirror
50 Years Later, MIT Looks Back At AI and Networking Pioneer Project MAC
v3rgEz (125380) writes "Fifty years ago, a major project that ultimately seeded much of today's computer technology was created at MIT: Project MAC, and the Multics operating system initiative within the project. Daniel Dern interviews some of the key figures involved in the pioneering project, looking at how one laboratory helped spawn Ethernet, AI, and dozens of tech companies and other innovations that took ideas from the lab to the personal computer."
Dude, what the hell is your problem? You think that other people can't tell that it's the same person making all these troll first posts lambasting "Space Nutters" or lampooning the impact on popular technology ascribed to the space race? Your shtick might have been funny many months ago, but now it's just lame -- and hopefully it's just a shtick, not some kind of mental illness that drives you to post like this again and again.
MIT Project MAC
Slashdot, fix the reply notifications... You won't get away with it...
Unix is not Multics (thus the name) and that is really all you need to know about Multics (except that a generation of MIT hackers cut their teeth in figuring out how to hack it).
Machine Aided Computation.
It's the typical Cambridge (Mass) conceit that if it doesn't happened within 50 miles of Kendall Square (or Harvard Square, depending on your persuasion) it doesn't count, and if it does, you should know about it.
You want the "forgot to take my meds this morning" thread.
Oh GOSH it would be BLASPHEMY to QUESTION how OTHERS DO IT
---
Multics implemented a single level store for data access, discarding the clear distinction between files (called segments in Multics) and process memory. The memory of a process consisted solely of segments which were mapped into its address space. To read or write to them, the process simply used normal CPU instructions, and the operating system took care of making sure that all the modifications were saved to disk. In POSIX terminology, it was as if every file was mmap()ed; however, in Multics there was no concept of process memory, separate from the memory used to hold mapped-in files, as Unix has. All memory in the system was part of some segment, which appeared in the file system; this included the temporary scratch memory of the process, its kernel stack, etc.
---
REALLY you might want to pull your head out of the sand you might learn something
Then why was it called God's Chosen Operating System? Huh, smartass?
Yes, I'm lampooning the simplified world-view of the space fanbois. Computing, electronics and technology in general are complex subjects, and many, many kinds of people worked on many kinds of technologies for many reasons, even in countries without a manned space program. Imagine that.
The people that absurdly oversimplify history into NASA=technology deserve as much "lambasting" as my keyboard allows (and /.'s stupid 10 posts a day AC limit.)
And if a knowledge of history is a mental illness, then stop the world, I wanna get off.
Sometimes I also make AC posts that mimic a Space Nutter's talking points and guess what? I can get +5 insightful quite rapidly.
There is something deeply wrong about how geeks view reality. Witness the absurd claims about 3D printing, or the dead-serious faith that the species has to get off this rock!
Youtube has many films from the era.
https://www.youtube.com/watch?...
Mr Corbato himself.
Mostly random stuff.
The stuttering black and white film with audio dropouts reminds me of the Dharma Initiative orientation films from Lost.
When our name is on the back of your car, we're behind you all the way!
Knowledge of history isn't a sign of mental illness at all, but an overwhelming urge to make others read your views on history even in discussions of another subject might be. The Slashdot submission did not claim any link between this technology and the space race, and the comments section could not have either since yours was the first post.
So, the other commenter's comparison of you to APK is apt. You are not educating anyone here on history, you are only making others concerned about your sanity.
Multics was amazing for its time. Then Honeywell took it on as a commercial product and didn't know how to sell it, or more like their sales people were clueless and didn't want it competing against their own home-grown crappy operating system and hardware. So Multics died.
In the minicomputer era, Prime Computers (a competitor to DEC) built an operating system that they called a "mini" Multics, because it used the same security ring idea, but it wasn't a tenth as good.
And I do myself remember Project MAC as I was a student at MIT in those days.
You can keep claiming your narrow alternative history all you like, does not make it any more accurate.
*shrug* I am under 40, never went to MIT or any of the other classic CS schools and I knew what this was. And when I do not recognize something, I google it.
Biz saw no need for computers at first. Market of six, etc. Govt drove the industry with orders during the 50s and 60s. NASA was like 4% of GDP and represented a big part of the orders. Conservative short-sighted biz would have been content to wait for centuries before responding to a pure market signal for miniturization. So yes, NASA played a large role.
Minsky's problem is that he just writes, doesn't program his ideas, which are a pretty good start: i.e., a society of minds or independent programs with different domains. The programs should communicate with each other and with the user through natural language, instead of arbitrary programmer-defined protocols.
It is the handle of a quite a prolific poster here. You can recognize his posts quite easily because of the lengths and massive amount of links.
Was quite happy making typewriters. Have you shown evidence that they were working on miniaturizing without government involvement. I think there were many factors involved in the technological advancements of the past 1/2 century or so and the space race (militarization and ICBMs included) did play a large role in that. Would we have gotten there without it? Probably. Would we have gotten there quicker or even as quickly? Probably not.
IBM happily took orders from the Nazi govt.
My truename appears in the list of "Multicians". I still have my copy of The Design of the Multics Operating System.
One of the things that Multics did better than anything since was a feature called dynamic linking. In Multics, linking to a DLL was done via a symbolic reference resolved at runtime, rather than a reference to an ordinal (as in Windows). The Multics file system allowed you to have multiple names on the same file. The combination of those two features resulted in the ability to hot-plug DLLs. Here's how:
1. You have a program which wants to take a sine function. It's got a link to "fortran.lib:sin" (no, the Multics syntax for the entry point was different, but you get the idea).
2. The fortran guru decides he wants to upgrade the fortran library while your program is running.
3. When your program first invoked the sin() function, the symbolic link is resolved to the existing fortran.lib DLL and it's loaded into your pdd (process space).
4. The fortran guru adds the name "fortran.lib_bak" to the existing library file.
5. The fortran guru creates his library as "fortran.lib_new".
6. The fortran guru moves the name "fortran.lib" from the old library to the new one.
7. Immediately, user programs which have not already linked to the old DLL will now link to the new one as symbolic references are encountered and resolved.
8. The fortran guru removes the name "fortran.lib_new" from his new library, which has no effect except to free up that name for use in a future upgrade.
9. Eventually, all user programs which referenced the old library finish and the old library can be deleted. Everyone now uses the new library,
When I first learned about this, I thought is was really cool. 35 years later, I still do.