BSDs Struggle With Open-Source Graphics Drivers

[Sergio]

I have in front of me at the moment, a copy of the famous "Chine Nual", dated 1984. The preface to this document is the VERY FIRST PUBLIC MENTION of the gnu project:

"I believe that the commercialization of computer software has harmed the spirit which enabled such systems to be developed. Now I am attempting to build a software-sharing movement to revive that spirit from near oblivion"

WHAT is the "spirit" that develops systems??? It is RESEARCH! HELLO! What else is it?

and DO YOU SEE YOUR ERROR? You see GNU as a software project. It's NOT a software project. It's a project to invigorate software design. It's a project to get people thinking, to stir their brain cells and be better. YOU think it's about the software.

And by the way, WHAT DO YOU THINK HURD IS FOR, ANYWAY?? Do you REALLY think the developers have any intention of "shipping a software product"??? Hurd is a learning experience, it's a RESEARCH PROJECT, it's a sandbox for new ideas and experiments. It spins off ideas like FUSE that get integrated into other things.

Obviously we have very different definitions of "research". The HURD is based on Mach, WHICH WAS A RESEARCH SYSTEM. Plan 9, MINIX, L4, Inferno, Amoeba... these are all research systems. The HURD does not, according to my notion of research, belong to this group. So no, the "spirit" of those systems (GNU) is NOT RESEARCH.

[LightBit]

All those people who moan and complain about Apple and what has happened to their products, you can point your finger right square at the BSD license and say, "SEE??? This is what you get when you put the BSD license on your code. Someone else will take it and polish it and put it in their deus ex machina and take all the credit, and leave you with nothing other than the bitter knowledge that you helped create the new Frankenstein monster.

GPL people are complaining about Apple taking BSD code. You know some people don't have problem with that (their code will stay open).
Maybe if TCP/IP wouldn't be BSD licensed, Microsoft would implement different protocol.

As a user I don't have any problem using BSD, GPL or any other open source licensed software.
But as programer I prefer BSD license in most cases (libraries and small utilites), because is simpler and more compatible.

Nobody forces you to release your code under BSD license or GPL.

[frantaylor]

There are people who complain, "the GPL is against freedom and totalitarian, because the GPL doesn't allow people to do whatever with the code! That's not real freedom!"

There are also people who argue, that "the laws that forbid slavery are against freedom and totalitarian, because these laws don't allow people to do whatever they want to other people! That's not real freedom!"

Both are equally right. Yes, sometimes we need laws that proximately limit someone's freedom to do some thing, in order to ultimately guarantee more freedom for more people. For example: In order to give people the freedom to not be exploited, we have laws that limit the freedom to enslave others. In order to give people the freedom to walk on the streets without getting assaulted, we have laws that limit the freedom to assault others. And the GPL works the same way. In order to grant the user of the software the four freedoms, it limits the freedoms of whoever wants to fork the software or contribute to it's development.

The BSD license on the other hand says "we don't want to limit anyone's freedom", so they ultimately end up ignoring everyone's freedom. What would happen if laws worked the same way as the BSD license? There would be no laws against slavery, because that would be limiting the freedom of the slaveowners. There would be no laws against assault, because that would be limiting the freedom to assault whoever you want. Thus, you would not have the freedom to not be exploited, nor would you have the freedom to walk on the streets unharmed.

All those people who moan and complain about Apple and what has happened to their products, you can point your finger right square at the BSD license and say, "SEE??? This is what you get when you put the BSD license on your code. Someone else will take it and polish it and put it in their deus ex machina and take all the credit, and leave you with nothing other than the bitter knowledge that you helped create the new Frankenstein monster.

[dee.]

There are people who complain, "the GPL is against freedom and totalitarian, because the GPL doesn't allow people to do whatever with the code! That's not real freedom!"

There are also people who argue, that "the laws that forbid slavery are against freedom and totalitarian, because these laws don't allow people to do whatever they want to other people! That's not real freedom!"

Both are equally right. Yes, sometimes we need laws that proximately limit someone's freedom to do some thing, in order to ultimately guarantee more freedom for more people. For example: In order to give people the freedom to not be exploited, we have laws that limit the freedom to enslave others. In order to give people the freedom to walk on the streets without getting assaulted, we have laws that limit the freedom to assault others. And the GPL works the same way. In order to grant the user of the software the four freedoms, it limits the freedoms of whoever wants to fork the software or contribute to it's development.

The BSD license on the other hand says "we don't want to limit anyone's freedom", so they ultimately end up ignoring everyone's freedom. What would happen if laws worked the same way as the BSD license? There would be no laws against slavery, because that would be limiting the freedom of the slaveowners. There would be no laws against assault, because that would be limiting the freedom to assault whoever you want. Thus, you would not have the freedom to not be exploited, nor would you have the freedom to walk on the streets unharmed.

[frantaylor]

the GNU system wasn't meant to be a research-level system

I have in front of me at the moment, a copy of the famous "Chine Nual", dated 1984. The preface to this document is the VERY FIRST PUBLIC MENTION of the gnu project:

"I believe that the commercialization of computer software has harmed the spirit which enabled such systems to be developed. Now I am attempting to build a software-sharing movement to revive that spirit from near oblivion"

WHAT is the "spirit" that develops systems??? It is RESEARCH! HELLO! What else is it?

and DO YOU SEE YOUR ERROR? You see GNU as a software project. It's NOT a software project. It's a project to invigorate software design. It's a project to get people thinking, to stir their brain cells and be better. YOU think it's about the software.

And by the way, WHAT DO YOU THINK HURD IS FOR, ANYWAY?? Do you REALLY think the developers have any intention of "shipping a software product"??? Hurd is a learning experience, it's a RESEARCH PROJECT, it's a sandbox for new ideas and experiments. It spins off ideas like FUSE that get integrated into other things.

[Sergio]

WOW how thick are you??? I am talking about a TEXT EDITOR! I am talking about MACHINES WITHOUT VIRTUAL MEMORY like PDP-11. Back in 1986 we were running UNIX and emacs and lots of other programs on PDP-11 systems with NO VIRTUAL MEMORY!!! OMG can you believe it!

I am talking about BUFFER MANAGEMENT TECHNIQUE which was NOT AVAILABLE in editors like VI at the time. If you tried to use vi to edit a file bigger than RAM, it would just FAIL with an out-of-memory error BECAUSE THERE WAS NO VIRTUAL MEMORY!!! However emacs was VERY CLEVER in this reduced memory situation and could edit this file.

You said that only Emacs let you edit a file bigger that physical memory. So, suppose there was no virtual memory (PDP-11 HAD virtual memory in the form of segments). So, did Emacs invented overlaying? or perhaps bank switching? No. So, Emacs implemented a form of overlaying that allowed to edit a file bigger than physical memory without virtual memory? WOW, WHAT A TOP-LEVEL RESEARCH! HOW COULD I MISS THAT!

How thick are you...

Any way, I admire and value what the GNU folks have accomplished, but my point was that the GNU system wasn't meant to be a research-level system; its purpose was to provide a free Unix. Ok, so they did interesting things, maybe I wouldn't call that research, maybe you would. But in that respect, BSD and GNU were fundamentally different in that BSD was born as a research system.

[frantaylor]

Are you talking about virtual memory? the same that was available since 1961 through the Atlas Supervisor? But virtual memory is completely transparent to applications, so yes, you could in theory edit a file bigger than the available physical space.

WOW how thick are you??? I am talking about a TEXT EDITOR! I am talking about MACHINES WITHOUT VIRTUAL MEMORY like PDP-11. Back in 1986 we were running UNIX and emacs and lots of other programs on PDP-11 systems with NO VIRTUAL MEMORY!!! OMG can you believe it!

I am talking about BUFFER MANAGEMENT TECHNIQUE which was NOT AVAILABLE in editors like VI at the time. If you tried to use vi to edit a file bigger than RAM, it would just FAIL with an out-of-memory error BECAUSE THERE WAS NO VIRTUAL MEMORY!!! However emacs was VERY CLEVER in this reduced memory situation and could edit this file.

By the way, I LOVE how you assume that virtual memory was everywhere in 1986! Back in that time, virtual memory was not considered to be a real solution to anything at all. In order to have virtual memory you needed to purchase additional hardware that cost even more than the CPU and used more power than the CPU. It inserted wait states into your memory fetch cycles and it slowed your computer way down. Virtual memory was considered to be the "lazy way out" for people who were too stupid to fix their software.

Back in those days of 4 MHz processors and 250 ns memory cycle times, you needed lots of cleverness and RESEARCH if you wanted your program to run in any kind of sane manner. You couldn't just wait until next year and buy a faster machine, progress was much slower then.

[Sergio]

anybody who thinks that gnu was not doing "research" they are clearly delusional

why don't you go back to 1986. Get yourself a nice sized text file, say a megabyte or two. Yeah, it's WAY bigger than your machine's available physical memory, back in 1986. Now I would like you to try to open and edit this document with ANY text editor program that was available in 1986.

you will find to your vast chagrin that emacs will be the ONLY editor that will edit this file that is bigger than your physical memory. tell us more about gnu does no research.

Are you talking about virtual memory? the same that was available since 1961 through the Atlas Supervisor? But virtual memory is completely transparent to applications, so yes, you could in theory edit a file bigger than the available physical space.

[Sergio]

So let's see here, the otherwise-unknown art of text buffer management that emacs brought to the table, ALL of the work on gcc, the innovations in bash, guile's advances in dynamic language design, these things are not research??? Remember that all this stuff happened back in the 80's when Apple IIs were the desktop computer and MS-DOS was "state of the art".

And GNU has never been academic? Are you talking about the project that was started in a university research laboratory? The project that was nutured for many years by a tenured college professor? The project which employs mostly college students to do the grunt work? The project whose results are used for research and education all over the globe? This project has NEVER been academic?

WOW and you expect anyone to take you seriously?


WOW

Didn't Vi already have that otherwise-unknown art of text buffer management? What research-level work in GCC? What research-level innovations in bash?
When I say research I mean: first, the purpose of the project and second, effectively the research-level things its development brought.
If GNU started at the University it doesn't necessarilly imply it was conceived for academic purposes; indeed it was not.
Rewriting user-space tools and a kernel isn't research either, even if done by students. Linux is also used as base system for doing research. This doesn't mean Linux is an academic project. The same goes for GNU.
But BSD was the system that connected to world; this was possible due to research.
If you consider the "innovations in bash" a research-level project, do you expect anyone to take you seriously?
WOW.

[frantaylor]

anybody who thinks that gnu was not doing "research" they are clearly delusional

why don't you go back to 1986. Get yourself a nice sized text file, say a megabyte or two. Yeah, it's WAY bigger than your machine's available physical memory, back in 1986. Now I would like you to try to open and edit this document with ANY text editor program that was available in 1986.

you will find to your vast chagrin that emacs will be the ONLY editor that will edit this file that is bigger than your physical memory. tell us more about gnu does no research.