Date: Sat, 4 Jan 2003 21:15:42 +0100 From: Cliff Sarginson <clsn@raggedclown.net> To: FreeBSD Chat <FreeBSD-chat@FreeBSD.org> Subject: In the land of the blind a one eyed man becomes king Message-ID: <20030104201542.GA10588@raggedclown.net> In-Reply-To: <4.3.2.7.2.20030104112345.02a48b70@localhost> References: <200212312041.gBVKfr183480@hokkshideh2.jetcafe.org> <3E120659.3D60EB30@mindspring.com> <20030101140530.GA11468@raggedclown.net> <4.3.2.7.2.20030104112345.02a48b70@localhost>
next in thread | previous in thread | raw e-mail | index | archive | help
I am moved to reply. On Sat, Jan 04, 2003 at 11:41:04AM -0700, Brett Glass wrote: > At 12:36 PM 1/1/2003, Terry Lambert wrote: > > >Seriously, though, there are a number of possible compilers, > >but FreeBSD keeps adding constructs and removing portability, > >and, in general, getting more and more GCC dependent, as time > >goes on (hmmm... RMS paying pwople off?...). > > No, it's a combination of two things. First of all, RMS and > the FSF engage in "embrace and extend" tactics (which they > condemn when Microsoft uses them.... Hmmm.). Make use of > any of the unique quirks of GCC, or its internal assembler > with its "reversed" syntax, and before long you literally > have to port your code to get it to compile on any other > C compiler. > Oh I think we know why this is matey. A Norwegian, I think he goes by the name of Linus (or is that the dog in .. Peanuts). He was a brilliant student. He took a small project, invented to teach students operating systems principles and created Linux. GCC has been intimately involved, because it contains (now) language constructs that are not found in "Muppets can Learn C in 24 Hours Books". It has things in it to make writing device drivers easier. I have managed to scrape a living for a lot of years as a "C" programmer, and sometime System Manager (I gave up the latter because the stress overomes the money). There is no "C" compiler. Not one in the land of buggery that actually conforms to the "The Whitebook". Either the pre-ANSI or post ANSI version. I have worked on "porting" systems, from one machine to another. I won't bore you by enumerating them. I counted them once. 12-15 versions of "C". Some of it was hard work. Don't talk to me about PRAGMAS .. > Secondly, the FreeBSD project is insufficiently vigilant regarding > the issue of dependency on GPLed code. Of the BSDs, FreeBSD > has the most GPLed code, and would be the most crippled if > that code were removed. This is sad, IMHO, because it prevents > FreeBSD from being truly free. Most of FreeBSD may be licensed > under the BSD License, but what good is that if one can't install > it on one's machine without bringing in the GPL? > The alternitve being ... ? > OpenBSD pretty much avoids the use of GPLed code except in the > toolchain. I'd like to see them take an even tougher stance, > and eject the GPL from their code base altogether. But while > they removed ipfilter over a much smaller licensing issue, > they have not done the same with GPLed code. Why? One can only > speculate that it's because GCC is now so dominant that > other compilers just aren't obtainable... and also because so > much of the third party software for OpenBSD is dependent > upon GCC as well. A vicious cycle that needs breaking. > OpenBSD. Open to whom ? > >In any case, a > >compiler is almost trivial; what's hard, and takes specialized > >knowledge, is optimizing, and code generation, for more than > >one CPU family. > > True. Is that true ? Have you tried writing one ? Come on,. make it public, > > >RMS' great contribution in this regard is *not* the compiler > >itself; what he contributed there is actually a mediocre set > >of code, that other people then worked on to turn it into what > >it is today. In fact, he had to eat crow on EGCS to get it to > >be only a version successor to GCC, rather than a competing > >project. So RMS's contribution was the project, itself: the > >difficulty of writing a compiler is actually irrelevent to the > >discussion. > No it is not. Write one, display the code, tell us it will compile anything written in "C". > Also true. It's amazing how often one hears RMS described as > a master coder, when in fact the original GCC was awful. > I have never seen RMS described so, He was just someone with an idea. > >It's amazing to me the number of people who claim to study the > >Open Source Software phenomenon, yet don't understand the basic > >principles through which it actually functions, well enough to > >start, or help start, a project and have it persist. > > It's simple: It's what I call the "Yosemite Sam Principle." > > Remember the classic cartoon in which Yosemite Sam attempts to get > Bugs Bunny to blow himself up by playing a booby-trapped piano? > > Sam challenges Bugs to play the "pie-anner," which is rigged to > explode when a particular key is pressed. Bugs plays the tune from > the sheet music, but keeps hitting a wrong note and doesn't set > off the explosive. Finally, the frustrated Sam leaps in front of > Bugs, yelling, "No, you stupid rabbit! Like this!" and plays the > tune correctly. > > The explosive goes off. > > This is one of the secrets of open source: One of the ways to get > people to do something right for you is to do it wrong yourself > and let other people's egos intervene. They'll soon do it better > than you could (though, in the case of GCC, not necessarily as > well as it could be done). > Nope. We are all imperfect. Even Richard Stallman, You can sleep with Bill Gates.. No-one is stoppimg you ... > --Brett Glass -- Regards Cliff Sarginson The Netherlands [ This mail has been checked as virus-free ] To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030104201542.GA10588>