OSDev.org

[Your operating system] failed simply because it fell short of being a compelling enough improvement on Unix to displace its ancestor. Compared to [your operating system], Unix creaks and clanks and has obvious rust spots, but it gets the job done well enough to hold its position. There is a lesson here for ambitious system architects: the most dangerous enemy of a better solution is an existing codebase that is just good enough. - Eric S. Raymond

Of course, that only works with Unix because Unix is more of an idea than a system (because many systems can be classified as it.) If there is any sufficiently successful system that has similarities to a *nix system, it then becomes a *nix system itself. The original Unix is hardly used anymore, but it would be difficult to say that Unix has died.

That statement is therefore self-proving.

No, actually there's quite some truth in it. You could just as well replace "Unix" with "Windows" in that quote, and it would still bear the same important message.

Architecture, when looked at in isolation, might win prices, but it doesn't win market share.

Basically, that statement essentially is a well known adage applied to computer science.

Don't change the winning team

That is probably why people don't want to use your stuff, better or not.

Yeah, but who exactly is the winner? The development team or the marketing department?

Sadly enough, it doesn't matter.

Personally, I would say Microsoft has proven beyond any doubt that all it needs is a good marketing / business department, the tech geeks be damned.

@NickJohnson: The original quote is between Plan 9 and UNIX, and actually means the unix operating system.

This has happened to many projects time and time again (as combuster shows).

Does this mean you can't make a difference? or make a change? - I think what it really shows is that you have to offer more than just technical superiority to users.

Mike

Well, if it literally means UNIX, then the fact that Linux (and to a lesser extent, BSD) have replaced nearly all traditional UNIX installations is also a problem. I wouldn't say that Linux had many technical advantages over UNIX, especially when it was new; the only special thing about it was that it was free software.

Linux is a kernel and a bunch of drivers, nothing more. Most of the software running on top of that kernel is designed to work just as well on any other Unix. Many titles even run just as well on MacOS or Cygwin.

Linux is a Unix, just like Solaris, IRIX, AIX, HP-UX, *BSD, whatever.

@Solar: I know: I said if the quote literally meant the original UNIX (as in, the original UNIX code, which is not in Linux or BSD.)

The first serious fork of Unix (BSD) happened in the mid-70ies. By that time, Raymond still did APL, Pascal and Lisp for Wharton School. His first Unix-related work started 1983... by which time there was also Xenix, SunOS, and System III / V to be considered. I guess by the time he was "authoritative" enough to be quoted, Linux was a reality.

Ah, the early days. I guess no-one can believe the breakneck speed at which things changed back then, unless he was there and witnessed it first-hand.

When the 70ies drew to a close, there was Unix on VAXens. Little else.

At 1980, the IBM PC and the ZX-81 were but one year away.

From there, one year till the Commodore 64 arrived (color! games!).

Three more years to the Atari ST and the Amiga 1000 (more color, great sound, multitasking!).

One year after that the Compaq Deskpro 386 (uh... yeah ).

And that's still four years before Windows 3.0 (uh... yeeaahhh... )...

And that's five years before Win95 (multitasking... hell yeah. ).

What's new today? We can carry our computers around, store more stuff and have our 3D framerates keep up with ever-more-demanding 3D engines yet still yearn for the good old days of M.U.L.E., The Last Ninja or Elite.

And that's 16 years after Win95...

I dont think a huge amount of things have changed for a long time. Computing has become a bit stale.

Software is like a gas - it expands until it fills all available space and resources.
How it does it is beyond me, but some of it don't do that.
Linux takes 2 seconds to boot up on 366 Celeron and same 2 seconds of quad-core i7.
But new versions of popular programs take seconds to load up on newest hardware and minutes on old, while having hardly any more features than the older versions which load for seconds on old hardware and instantly on new.
Something is wrong in this world.

Solar wrote:Ah, the early days. I guess no-one can believe the breakneck speed at which things changed back then, unless he was there and witnessed it first-hand.

When the 70ies drew to a close, there was Unix on VAXens. Little else.

in the mid 70's ^.... by the close of the 70's CP/M and it's multitasking / multi-cpu, twins MP/M and Turbodos were the new up and comers. There were also dozens of totally incompatible computer systems on the market with totally incompatible custom OS's

But it was never really about the OS and only in a small way was it ever about the hardware. In the end it's about what the 'customer' thinks the computer can do and what programs it can run.... Realistically the hot new OS's are Android and iPhone OS. Even there it's still mostly about the aps. To replace an old OS you would need to be able to run programs or hardware that people want to use and that the old OS's don't. That I think is the only thing I that would make it 'better' in the public eye.

just my two cents

roboman wrote:

Solar wrote:Ah, the early days. I guess no-one can believe the breakneck speed at which things changed back then, unless he was there and witnessed it first-hand.

When the 70ies drew to a close, there was Unix on VAXens. Little else.

in the mid 70's ^.... by the close of the 70's CP/M and it's multitasking / multi-cpu, twins MP/M and Turbodos were the new up and comers. There were also dozens of totally incompatible computer systems on the market with totally incompatible custom OS's

But it was never really about the OS and only in a small way was it ever about the hardware. In the end it's about what the 'customer' thinks the computer can do and what programs it can run.... Realistically the hot new OS's are Android and iPhone OS. Even there it's still mostly about the aps. To replace an old OS you would need to be able to run programs or hardware that people want to use and that the old OS's don't. That I think is the only thing I that would make it 'better' in the public eye.

just my two cents

I second that to a degree.. but i disagree to your solution of how to replace OS's.

I think someone needs to build an OS that realises the 4 'customers', to clearly define their rolls, and to provide them with a simple and painless way for them to work.

Software Engineers - I believe hardware and software demands are so high, designing algorithms and optimisations really needs to stay out of programming and should be done by academics or dedicated people.

Programmers - Should be left to make software aimed at a task, they should be left to focus on application specific details and not have to worry about what is the best way to store their data or which type of collection would be best for them.

Administrators - Should be able to customise applications for the users needs. Such as passwords, company customisations etc.

Users - Software should be suitable for the job at hand.

I really think this is lacking in computing at the moment, and would be a great change for all.

Mike