Saving Unix Heritage, One Kernel At a Time

coondoggie writes "In this, its 40th year of operating system life, some Unix stalwarts are trying to resurrect its past. That is, they are taking on the unenviable and difficult job of restoring to their former glory old Unix software artifacts such as early Unix kernels, compilers and other important historical source code pieces. In a paper to be presented at next week's Usenix show, Warren Toomey of the Bond School of IT is expected to detail restoration work being done on four key Unix software artifacts all from the early 1970s — Nsys, 1st edition Unix kernel, 1st and 2nd edition binaries and early C compilers. In his paper, Toomey states that while the history of Unix has been well-documented, there was a time when the actual artifacts of early Unix development were in danger of being lost forever."

Comments

Usenix attendees.....

[KingPin27]

In a paper to be presented at next week's Usenix show, Warren Toomey of the Bond School of IT is expected to detail restoration work being done on four key Unix software artifacts all from the early 1970s

Afterwards atendees will be ushered to the dining hall for a fine serving of raisins, prune juice, and Oxygen treatments.
St. John's ambulance will also be on site to assist with attendees suffering with various age related ailments such as broken hips and arthritis.

Re:Why?

[Darkness404]

Unix really was one of the few programs to determine the fate of an entire industry. Every modern OS can trace back to Unix in some way or form. Keeping the history of Unix especially the early releases and plans can help better document the historical software.

Re:Why?

[rcamans]

Only if you know it is going in a straight line.

Re:Why?

[Anonymous Coward]

VMS? Windows? ReactOS? Plan9? QNX? Tron? zOS?

Re:Why?

[ArsonSmith]

Those that don't learn UNIX are doomed to reinvent it. Poorly.

Re:Why?

[oldspewey]

Maybe they're looking for the stolen SCO code in Linux?

Re:Why?

[Sponge Bath]

Tron?

He fights for the users.

Re:Why?

[qortra]

Not true - for example, if shown the first 3 cycles of a sinusoidal wave, I'm sure you could predict the next cycle. There are lots of non-linear numeric sequences that allow for relatively accurate predictions.

Obviously, history is more complicated, but idea is the same.

Re:Why?

[Culture20]

Every modern OS can trace back to Unix in some way or form.

VMS? Windows? ReactOS? Plan9? QNX? Tron? zOS?

I wouldn't call VMS modern...
Windows: take a look in c:\windows\system32\drivers\etc\ some time. See any Unix style influences?
The others I have no clue about.

Worse is better

[Anonymous Coward]

Despite what many a slashdot crowd might think, UNIX isn't exactly an elixir from the Gods. UNIX, Microsoft Windows and Intel x86 are living proofs that the best / most innovative technology doesn't necessarily have to win. Check Out: http://www.dreamsongs.com/WorseIsBetter.html.

Re:Worse is better

[LWATCDR]

The fact that MS-DOS outsold the Amiga and Atari ST is proof that best doesn't always win. The x86 is a great example as well. The 68k chip was a much better CPU than the 8088,8086, and even the 80286. Only when the 386 hit the market did Intel really have a CPU that wasn't a freaking nightmare.
Another example is PHP. Good grief $A[1]==$A['1'], that is just wrong.
PHP, Windows, x86, and so much of what we live with are all examples of good enough. Not great but good enough.

Re:Worse is better

[Teckla]

The fact that MS-DOS outsold the Amiga and Atari ST is proof that best doesn't always win. The x86 is a great example as well. The 68k chip was a much better CPU than the 8088,8086, and even the 80286.

The 68000 didn't have a built-in MMU. You could run an OS with process isolation (a requirement for a safe multi-user OS) on the 80286. You could not do that with a 68000 (unless you added a separate MMU; the 68881 maybe?).

Re:Worse is better

[LWATCDR]

The memory protection on the 286 was terrible. It still used segments and wasn't a flat address space.
I think the the 68010 came out around the time of 286 and you could add a real mmu that supported paging and had a flat address space. Also the 68020 came out during the time of the 286 so you could make the jump to true 32 bit.
Of course the vast majority of 286s where used to run DOS so it's memory protection just wasn't that important. BTW the 68881 was an FPU the Motorola 68841 or 68851 where the MMUs.

Re:Worse is better

[Darinbob]

That is a bad comparison, comparing early generation chips to a later one. The 8086, 8088, and 80186 didn't have built-in MMUs either! Compare the 68020 to the 80286 for a more reasonable comparison.

Memory protection was not at all a design factor for DOS or even early versions of Windows. Memory protection didn't really take off in the PC world until the 80386, which was after the 68020 was being used to run UNIX.

The two competing families didn't release new CPUs at the same time, so it's difficult to say which CPU should compare to which; they tended to leap-frog, as in 68020 being better than 80286, then 80386 being better a year or two later, then 68030 being better a year or two after that, etc.

This is in terms of MMU that is, the overall architecture of the 680x0 family was always much cleaner than the 80x86 family in my view. When you're comparing MS-DOS to Atari ST or Amiga, the 68000 is a clear winner. The reason that IBM did not go with that family for the first PC was due to the availability of support chips as I understand it.

Re:Why?

[ArsonSmith]

I studied this and found that June 4th, 2014 will be the start of the year of the Linux Desktop.

Re:Why?

[killmofasta]

Overgeneralization.

Old source code gives us ideas, like looking at the design philosophy behind the code, and the ultimate operation of the software. These are actually *priceless* artifacts, and since they are mostly digital ( reserive the right for first pun... they are 'Digital' ), the study and the disemination of the early code is of extrodinary value to coders and software architects.

Of course its also invaluable to have their nemisises Multics and VMS alos preserved. I personally got an enourmous amount of respect for K&R reading the source code for the kernel (the V4), and the proto compiler. K&R, and the linux/GNU write well, wereas their MS counterparts wirte pretty crappy stuff.  I would also venture to guess that the code alone can serve as an example of how to write code.

I will look forward to taking a detailed 'History of the UNIX Kernel' class in the near future.

History is history

[Kupfernigk]

Is there really any useful purpose to decoding Sumerian clay tablets, or analysing dockyard records from the 18th Century? One of the things that differentiates civilised human beings from all other living things on this planet is that we study history and preserve things from the past. Perhaps it just doesn't need justification, it is part of what we are.

What, no documentation?

[bzzfzz]

From TFA: "documentation is missing or incomplete, source code is missing leaving only the binary executables, or conversely the source exists but the compilation tools to reconstruct the executables are missing."

Sounds like any number of projects I've had the pleasure of working on over the years.

Careful...

[damn_registrars]

Those that don't learn UNIX are doomed to reinvent it. Poorly.

One may interpret that saying as someone trying to incite a Linux / BSD war. We lost good men from both sides the last time that happened...

Re:Which "Unix" are they talking about?

[Anonymous Coward]

RTFA or even the summary per chance?

SIMH

[wandazulu]

SIMH is a hardware emulator for a lot of the machines Unix ran on (PDP-8, PDP-11, etc.). They also have some original Unix versions along with some other software for the other hardware they support.

I have run Unix V5 on a SIMH-based PDP-11, and it worked well, though it was strange to realize how fast it was running, in emulation, on a machine 1/16 its original size (Mac laptop).

Re:Why?

[Darkness404]

I don't know that much about VMS, but according to the Wikipedia page it uses Unix-y things such as X11, but its more separate than Unix being that it is little-used and totally proprietary to HP. Windows NT's kernel design was inspired by Unix, and it used (uses?) the BSD networking stack for TC/IP. If you look at NT you will notice a lot of similarities in the NT design that were first introduced in Unix. ReactOS uses a lot of source code from Unix programs in order to replicate the Windows functionality such as WINE. And are you kidding me about Plan9? That was inspired in the extreme by Unix and was meant to fix the flaws Unix had, if there was no Unix there would be no Plan 9. On the Wikipedia page for QNX at the top are

QNX (pronounced /kju n ks/ or /kju nks/) is a commercial Unix-like real-time operating system, aimed primarily at the embedded systems market

I would call that Unix. I can't find much info on TRON but it seems to be a not very modern OS in terms of design. zOS works on one processor architecture that is proprietary to IBM to run on mainframes, as such it isn't exactly a general use OS and you could probably find some Unix in it if you looked hard enough.

Unix as an "Idea Mine"

[mpapet]

Very often the technically 'best' implementation doesn't win and I'd like to see those stories from inside Unix. For me, that's a more interesting angle than just version/feature stories.

how about curses and text games?

[roc97007]

rogue, sail, wump, search (you have crashed into a planet), battlestar (in the closet is a kingly robe), mazewars, that mazewars-like curses game who's name escapes me, with a variety of weapons (satchel bomb... oooo...) that had destructable maze walls.

There are a variety of Rogue-like games out there that have been ported to current platforms, but the other ones, especially sail, search and mazewars, I haven't seen in years and years. These games were arguably part of our early Unix heritage because they enticed people to get a login and explore the OS, and for many of us (myself included) they were our motivation learn how to write termcaps for obscure terminals and emulators (the acid test was if Rogue would render correctly), learn programming to fix and enhance the games, and earn root access to do installs and fix permission issues.

Multi-user Unix games like sail and mazewars helped spread the Unix word because we were always trying to entice others to get a login so we could play with them. People with early PC experience couldn't even conceive of multi-user games.

Re:how about curses and text games?

[Tetsujin]

There are a variety of Rogue-like games out there that have been ported to current platforms, but the other ones, especially sail, search and mazewars, I haven't seen in years and years.

sail, at least, is part of the "bsdgames" package on Debian.

Re:Why?

[Tetsujin]

Are those Unix influences, or influences from an OS predating Unix? i.e. the same source that caused Unix to adopt that style? You make it sound like there was nothing before Unix.

Almost nothing. Unix was created on The Second Day.

Re:Why?

[LWATCDR]

VMS came after Unix so calling it a not very modern OS is kind of odd. X-Windows was never supposed to be Unix specific in fact there was even a version for DeskView way back when.
WindowsNT really owes more to VMS than Unix the chief architect came from Digital. Tron also came after Unix as well. zOS is still an extremely important OS so just because it only runs on one CPU I wouldn't just thow it away as well. As to finding some Unix in zOS frankly you would probably find more zOS in modern Unix than the other way around. IBM really did pretty much invent everything that Bell Labs did not and they where their first.
I really am not fond of the the write up about QNX. It is Unix like in someways and shares an API with it. but QNX is a micro kernel RTOS.
And lets be very honest. Unix came from Multics. Every OS has built on and taken ideas from other OS's. None of them is the pure root source.
I am a big Linux fan but I often wonder if we are too willing to keep Unix as our foundation. BeOS was a clean new OS and while I have never used it wonder if it may not be better than Linux and WindowsNT. It did some amazing things back in the day.

Re:Which "Unix" are they talking about?

[Chris Mattern]

Original flavor. All the stuff talked about in this article comes from before UNIX split into its hundreds of variants. In fact, these are so early that they come from before UNIX escaped out of Bell Labs. UNIX didn't start splitting into different flavors until about Versions 4 and 5.

Re:Why?

[Anonymous Coward]

Tron is the most used OS in the world.

It was invented in Japan and is an embedded device OS.

Re:Why?

[jdbausch]

Is there really any useful purpose to be served by dredging this up? Don't these guys have anything better to do?

yeah! FUCK HISTORY!

Re:Why?

[VAXcat]

Are you for real? Windows NT's kernel was practicaly a copy of VMS, it wasn't "inspired by UNIX". VMS was based on RSX. RSX came to be at approximately the same time as UNIX, circa 1970. RSX and VMS were influenced by UNIX allright, in a negative way - if UNIX did something, RSX and VMS definitely wanted to do it some other way.

Easy to use Windows SIMH packages

[JasonStevens]

I'm glad this is getting some exposure. I know that Warren & co worked hard to get this ancient UNIX not only in a working state, but also he is the one responsible for pushing SCO with the oldSCO source license, and played a hand in getting Research UNIX 1-7 & 32v under a BSD style license, thus setting the foundation of UNIX free. Now SIMH may not be the 'friendliest' software out there for a new user to get used to, so I've done my part in making it a little more accessible. On the sourceforge project https://sourceforge.net/projects/bsd42 I've created Windows installable versions of the 4BSD stuff, 32v and UNIX v1. I do plan to add all the other research versions, along with a new build of RENO that doesn't need 1.8GB... Anyways try them out! the 4BSD stuff has TCP/IP along with a SLiRP hack it can connect to the internet immediately! IRC/Lynx/GCC work great on the Uwisc 4.3 BSD build. Ok that being said, there is a repository of SIMH binaries on https://sourceforge.net/projects/simh , and the MS-DOS build includes some small 'bootstrap' versions of various OS's including v1 UNIX on the PDP-11 simulator. The bar to trying this stuff is a lot lower then you may have guessed, and I'd encourage any fan of UNIX to really check it out.

Oh boy

[FranTaylor]

Someone will trot out a copy of the Morris worm and we can relive history all over again.

WHAT'S WRONG WITH THE SLASHCODE?!?!

[ObsessiveMathsFreak]

OK seriously, the above post is pretty screwed up in Firefox. I've got floating tab bars or something all over the post as well as throughout the thread and the tt font is coming out at 16 point or in some very large font.

These css screw ups have been happening a lot lately. Then again I am using the older (and better) comment system.

paper and program

[adelporto]

Paper here: http://www.usenix.org/events/usenix09/tech/full_papers/toomey/toomey.pdf

Program here: http://www.usenix.org/events/usenix09/tech/

Yes, I work for USENIX.

Re:paper and program

[adelporto]

The paper is free next week when it becomes published.

Re:I do the same thing with used condoms

[NotBornYesterday]

Sadly, my not getting the joke rarely has anything to do with the quality of the joke itself.

Re:Why?

[morgan_greywolf]

Old source code gives us ideas,

Like "WTF did they do here?!", "Why did they do it THAT way?!" and "That has got to be the ugliest kludge I've ever seen!"

Of course its also invaluable to have their nemisises Multics and VMS alos preserved.

Multics was hardly a 'nemesis' of Unix. Multics was basically dead when Unix arrived; its death inspired Ken Thompson and company to work on Unix.

Re:Why?

[Kidbro]

Is there really any useful purpose to be served by dredging this up?

No. There is nothing useful to be learned from history.
Close your eyes, put the pedal to the metal, and assume that whatever you're doing, it's the right thing, and that nobody has ever tried it before.

Re:Why? (VMS)

[uassholes]

Here is something that might help. Excerpt: "And now...the rest of the story: I'll take you on a short tour of NT's lineage, which leads back to Digital and its VMS OS. Most of NT's lead developers, including VMS's chief architect, came from Digital, and their background heavily influenced NT's development. After I talk about NT's roots, I'll discuss the more-than-coincidental similarities between NT and VMS, and how Digital reacted to NT's release. . ." Link: http://windowsitpro.com/Articles/Index.cfm?ArticleID=4494

Re:Why?

[inamorty]

pitty the world will end in 2012 :(

Re:I love the smell of analogies in the morning!

[Tetsujin]

Restoring old cars saves energy and is environmentally friendly. I bought a used 1992 F250 needing work

No no no no no... I mean old cars...

keep regressing

[thethibs]

Actually, all these OS' lead back to the Berkeley Timesharing System (1964). So do many of the relevant people.

Re:Why?

[david_thornley]

At that time, lots of languages had short variable names. It wasn't seen as all that much of a big deal until later.

It was around this time, BTW, that I thought getting a 24x80 synchronous terminal on my desk was cool. (Synchronous = sends and receives in sort of a batch mode, asynchronous = character by character.)