| |||||||||||||
Phantom menace? Threat that, so far, has not existed? Sorry - you are DEAD wrong.Phantom menace? Threat that, so far, has not existed? Sorry - you are DEAD wrong.Posted Oct 4, 2008 7:25 UTC (Sat) by khim (subscriber, #9252)In reply to: GPLv3 anti-DRM clause by drag Parent article: Plugging into GCC
Aha. So NOW we are at stage where we are ignoring history if it does not suit us, right? Menace was clearly not phantom and threat was quite real - and you can read about this story here. These two sentences pretty much invalidate all your affectations: Why do we have a free C++ compiler? Only because the GNU GPL said it had to be free. That's pretty big accomplishment of all these "draconian restrictions" if you ask me. Of course that was then and we are talking about now, but... those who cannot remember the past are condemned to repeat it so we can not ignore the past and do everything you can to facilitate and create the best, most flexible, and most beneficial compiler you can and IF proprietary software becomes a problem THEN do something to try to deal with. Because the proprietary software WAS and IS the threat. Quite real and tangible. Yes, some companies like to play nice with free software (IBM, Google, etc)... as long as free software benefits them. Once that's not the case... all bets are off. Now, does it mean GCC should not implement plugins system? Of course not: it's usable for free software too. But do that and pretend that "proprietary threat" does not exist... it's just folly.
(Log in to post comments)
Phantom menace? Threat that, so far, has not existed? Sorry - you are DEAD wrong. Posted Oct 4, 2008 8:41 UTC (Sat) by drag (subscriber, #31333) [Link]
The license worked then... so it won't work now?
Look; There is a reason why people pored work into LLVM instead of improving GCC.
If you want people to use a GPL license and create open source software, making it impossible to do the work they need to get done is a shitty way of accomplishing that goal.
Have you actually READ the article? Posted Oct 4, 2008 10:16 UTC (Sat) by khim (subscriber, #9252) [Link] The license worked then... so it won't work now? Argh. WHY the license worked then? Apple and MCC specifically asked "can we distribute our front-ends as proprietary plugin to GCC" and RMS said "no" - after that GCC got C++ and ObjectiveC. And this worked because there were no simple way to write proprietary plugin and claim it's "independent work" (you were and still are pretty much forced to use large chunks of GCC code to interact with backend). Look; There is a reason why people pored work into LLVM instead of improving GCC.Yes and there are reason why Unix went nowhere while Linux took it's place. BSD went from being "something far beyond Linux" to "roughly equal, sometimes better, but often much worse". Yes, Apache/BSD licenses attract more developers and they do more work which is then wasted when companies go out of business. NetBSD tale should teach us something. The question today is: is GPLv3 (which makes it impossible for proprietary vendors to ship GCC with proprietary extensions - they can only ship extensions without appropriate copy of GCC) enough to prevent balcanization of GCC or not? If you want people to use a GPL license and create open source software, making it impossible to do the work they need to get done is a shitty way of accomplishing that goal.The question is, of course, not so simple. It's more like: do people want to create free software or do they want to create proprietary software, how many of the second camp can be pressured to first camp and how many of guys from first camp will go away in frustration if plugins will not be implemented? Answers to these questions are not easily obtainable - we can only guess...
Have you actually READ the article? Posted Oct 4, 2008 11:37 UTC (Sat) by Frej (subscriber, #4165) [Link]
Yes and there are reason why Unix went nowhere while Linux took it's place. BSD went from being "something far beyond Linux" to "roughly equal, sometimes better, but often much worse". Yes, Apache/BSD licenses attract more developers and they do more work which is then wasted when companies go out of business. NetBSD tale should teach us something. <snip>
That's true, there is no life time guarantee that someday you will be on your own using llvm when rest went closed. But in reality llvm+clang looks like a great piece of software (just avoiding the platform buzzword...) that can support existing C and C++ code...For now it just appears to be better software. Integrating clang with stuff like anjuta or emacs would finally bring actual code editors instead of text editors. ... ok just rambling now; sorry ;)
Have you actually READ the article? Posted Oct 19, 2008 1:55 UTC (Sun) by robert_s (subscriber, #42402) [Link]
"But in reality llvm+clang looks like a great piece of software (just avoiding the platform buzzword...) that can support existing C and C++ code...For now it just appears to be better software."
'For now' it doesn't support c++.
Have you actually READ the article? Posted Feb 5, 2009 14:22 UTC (Thu) by trasz (guest, #45786) [Link]
> Argh. WHY the license worked then?
It didn't. Marketing worked.
|
|||||||||||||