Is it correct to say that Unix is still alive?

[Remington]

The original Unix was created by AT&T Bell Labs. AT&T is a for-profit organization. Some people consider it to be one of the most most evil ones ever in the US economic system, a monopoly that for a long time had no equal. I see it more nuanced.

If it wasn't for AT&T System V then none of OSs that we have witnessed in our lifetime will exist today and we probably still be using landline phone. iOS and Android both are UNIX derivatives. It's true that AT&T was notorious company to begin with but UNIX changed everything technologically and we can thank AT&T for that.

The Unix certification process has one purpose: To determine whether an OS meets the interface standards to be interoperable with another set of OSes. I keep reminding people that this has nothing to do with the question whether the OS contains source code from the original Dennis&Ken prototype; as an example, IBM's z/OS has the Unix certification, and its heritage is as a mainframe OS (which predates Unix by about a decade or two).

Who determines what is UNIX standard and the requirements to pass the test to be certified? Again, they are in position to create or remove any requirements or specifications. If they certified an OS to UNIX and later they changed the requirements which they decertify the OS. That's why they don't have absolutely authority on which OS are UNIX or not. FreeBSD is UNIX based. So is OpenBSD, NetBSD, Solaris but they're all different. So what gives OpenGroup the authority to say which OS is UNIX and what's not? It has become so ambiguous as to what is defined as true UNIX. Alot has changed in last 36 years.

Have you tried using a mac at the bare OS level, without the GUI, just with C code, a compiler, and the standard system calls and libraries? It is Unix, in the sense of that everything one expects to work on Unix will work there too.

Have you tried installing macOS on an ordinary PC without using a bootloader? It's not true UNIX. Like I said, it's very heavily modified.

If you are trying to insinuate that Apple illegally bribed the OpenGroup to get the certification, that is not only complete nonsense, it is also libel and slander. Please stop making groundless accusations.

I didn't say that, again, OpenGroup is in position to change the rules anytime and OpenGroup and Apple was involved in lawsuits before.

Can you explain more about that? I must have forgotten about that lawsuit.

I meant that Open Group sued Apple because Apple used 'UNIX-based' and Open Group didn't like anyone using the word 'UNIX' without passing their certification process. Apple counter sued and claimed that UNIX is generic since its changed hands so many times, sued, counter sued, source codes changed, diverged or cloned into many different OS, etc.

Also AT&T sued BSD which resulted in settlement and BSD deleted 2/3 of AT&T source codes.

But as a reminder: Neither FreeBSD nor MacOS contain a single line of source from the original Bell Labs (a.k.a. Dennis&Ken) Unix.

If that's the case then all UNIX variants are not true UNIX.

 

[ralphbsz]

If it wasn't for AT&T System V then none of OSs that we have witnessed in our lifetime will exist today and we probably still be using landline phone.

No, there was lots of computer work before, in parallel with, and after Unix. There was multics, there was a whole slew of interactive OSes on mainframe (ever used CICS and TSO?), there was RSX-11 and VMS (which in many ways was a better Unix), there were a whole slew of research OSes in the 60s and 70s which are mostly forgotten today. And in the minicomputer world, there was a lot of interesting stuff, like RDOS (Data General), and my favorite perverse one, PrimOS (the only OS I know of that was written in Fortran), plus AS/400 (where every object is a memory address, sort of the polar opposite of Plan9, where every object is a file name), and S/38 (where everything was written around the relational database).

So what gives OpenGroup the authority to say which OS is UNIX and what's not? It has become so ambiguous as to what is defined as true UNIX.

There are may ways to define what the word Unix means. You are pushing one particular worldview. The Open Group has another one. I bet that there are many others. This is not mathematics, where everything needs to be clearly defined in axioms, and then all knowledge can be derived in the form of lemmas and theorems. Natural human language is more ambiguous, as is the term Unix.

Have you tried installing macOS on an ordinary PC without using a bootloader? It's not true UNIX.

That's not factually correct. MacOS contains Unix (in the sense of: it can do all the things other Unixes can do, for example compile and run any program that is written to the API standards), but it can also do other things (like serve the needs of the GUI, and restrict itself to run only on Apple-certified hardware, I know with hacky exceptions). What you are saying is like: Birds are not like animals, because animals can't fly, whereas birds can.

Also AT&T sued BSD which resulted in settlement and BSD deleted 2/3 of AT&T source codes.

That is a massive distortion of what really happened. To begin with, there is no way to sue "BSD" in the modern sense of the world. Instead, what was sued was first BSDi (a small commercial company, long since defunct, which was distributing an enhanced version of the Berkeley public release), and later the University of California (not specifically Berkeley, and in particular not the CSRG at Berkeley). Also, it wasn't AT&T who sued. Instead, USL sued first, but in the middle of the suit USL changed from being owned by AT&T to being owned by Novell and having nothing to do with AT&T. Second, the removal of all AT&T source code happened before the lawsuit, in the case of Net/1 long before the lawsuit. Third, in the middle of it all the University sued USL, claiming that they had illegally taken networking code from BSD and used it in their product, without the necessary attribution (BSD is freely distributable, and already was back then, but you can't remove the copyright notice).

The claim that what was removed was "2/3 of the source code" is also wrong; if you talk to people who were there (like Kirk McD or Lynn/Bill J), it was a relatively small effort to remove AT&T-licensed code in the early 90s, since so much had already been replaced in the CSRG at Berkeley. You have to remember that OSes used to be much smaller; the idea of dozens of millions of lines of source would have been laughable back then.

If that's the case then all UNIX variants are not true UNIX.

I keep bringing up AIX: AFAIK it contains AT&T licensed source code that does go back to Dennis and Ken. So this statement seems also false.

 

[CraigHB]

This thread prompted me to go back and read about the breakup of AT&T. Learned some things there. It's interesting the patent system allowed them to become a monopoly, but then then if you get too big that's not okay either.

Anyway this thread does make me think about what Unix really is. More of an idea or design philosophy than an actual product then. Still the idea is nebulous, what is it that defines Unix, is it POSIX or System V?

I thought one of the authors of the original AT&T Unix spent some time at Berkeley working on BSD. Wouldn't that make BSD a product of the original design? In that case it ~is~ Unix rather than a clone or a "Unix like" system.

 

[bookwormep]

Side note: You should listen to attorneys argue the definitions of words and semantics. This seems all too familiar.

 

[sidetone]

It is arguably Unix, but if it calls itself Unix, it gets sued, because Unix (tm) a trademark.

 

[ralphbsz]

It's interesting the patent system allowed them to become a monopoly,

Patents are actually a legally created and legally protected monopoly. Once you have a patent on something, the intent is that you could be the only one that can benefit from exploiting the idea, if you so wish.

In reality, a patent is nothing but a license to sue. A patent is worthless, unless you're willing to enforce it. The court system will help you with that, but to use the court system, you first have to front lots of money. In the case of patent lawsuits, millions of $, sometimes dozens of millions.

Anyway this thread does make me think about what Unix really is.

There are many different meaning of the word Unix. You picked one:

More of an idea or design philosophy than an actual product then.

There are many others. Two were discussed in this thread: Something that contains code originally written by Dennis & Ken. And: something that passes the Unix conformance test that the OpenGroup administers. All these definitions are sensible, and in many cases they overlap. But in some they don't.

People tend to think that words have exactly one meaning. This is nonsense. It kind of works in math, where one carefully defines things (an integer is ... and so on). Even there, it leads eventually to contradictions (Goedel, Russell, all that). In natural language (human language), words have many meanings. Claiming that "If it is Unix, it must follow a certain design philosophy" or such is just nonsense.[/quote]

I thought one of the authors of the original AT&T Unix spent some time at Berkeley working on BSD. Wouldn't that make BSD a product of the original design?

Ken Thompson took a sabbatical (half a hear or a year?) from Bell labs and spent it as Berkeley. I honestly don't know whether he spent more time hacking on Unix or teaching CS classes. I think some of Ken's family (perhaps his children) live in the bay area, so it may have been partially a family visit. But the Berkeley CSRG had a complete copy of the AT&T source code, and access to talk to the original authors (for example at conferences. And when BSD started, it was (by definition) 100% AT&T and zero Berkeley, while towards the end it was nearly completely non-AT&T code. So it was clearly heavily influenced by the original design and philosophy.

 

[CraigHB]

Claiming that "If it is Unix, it must follow a certain design philosophy" or such is just nonsense.

One of the things I really like about FreeBSD is I know what to expect from version to version. Saying Unix has no design philosophy means there's no standard. Things could be changed wildly and still be Unix. Though I know that's not the case for FreeBSD since it tries to conform to the principle of least astonishment. That's actually a big comfort for me. I really hate it when software is radically changed for seemingly no good reason, just ruins my day every time.