Re: BSD Licensed C++ compiler
From: toby (toby_at_telegraphics.com.au)
Date: 10 Nov 2005 10:21:34 -0800
John S. Dyson wrote:
> In article <email@example.com>,
> "toby" <firstname.lastname@example.org> writes:
> > John S. Dyson wrote:
> >> In article <email@example.com>,
> >> "toby" <firstname.lastname@example.org> writes:
> >> >
> >> > John S. Dyson wrote:
> >> >> In article <email@example.com>,
> >> >> "toby" <firstname.lastname@example.org> writes:
> >> >> >
> >> >> > Casper H.S. Dik wrote:
> >> >> >> "Karen Hill" <email@example.com> writes:
> >> >> >>
> >> >> >> >I know SUN has a modern C++ compiler available but AFAIK you can only
> >> >> >> >use it to compile OpenSolaris. Otherwise you must pay thousands for
> >> >> >> >one.
> >> >> >>
> >> >> >> No, the only requirement is that you are registered at the OpenSolaris
> >> >> >> website; there are no restrictions on its use as you'll (necessarily)
> >> >> >> be developing Solaris software with it.
> >> >> >>
> >> >> >> I think the main reason there's no BSD C/C++ compiler is that the
> >> >> >> GNU "C/C++" compiler is "good enough".
> >> >> >
> >> >> > Also because the GPL imposes no conditions on software you build *with*
> >> >> > it.
> >> >> >
> >> >> In this way, the GPL is little different than REASONABLE license terms of
> >> >> proprietary compilers. (Some licenses are horrible, but considering the
> >> >> GPLed compilers to be very similar to proprietary will work for most of
> >> >> us.)
> >> >
> >> > I also saw your earlier swipe at the GPL above.
> >> >
> >> If open and constructive discussion is considered a 'swipe', then perhaps
> >> the biased one is you.
> >> >
> >> > I call total BS.
> >> >
> >> Truth as I have stated is NOT BS. Fear of open, honest and accurate
> >> criticism is sad.
> > I'm not afraid of it; I'm just saying I think you're wrong. I also
> > asked for you to explain why you think you're right - see above.
> (Note that your rude and abusive language didn't encourage me to answer
> your comments. However, you were a little less rude, using a little
> less rude and unnecessarily abusive language, and so I'll answer you.)
Both of my posts were based on a foolish misreading of your position. I
> I truly didn't mean to insult your gender identity (or cause you similar
> difficulty in other realms.) This is not meant to be personal, and
> I am mostly suggesting that throwing caution to the wind, making any
> assumptions based upon the terminology (and redefinitions) of the term
> 'free' is not a good idea.
> MULTI-PAGE, MULTI-PARGRAPH licenses of so-called free software
> are very clearly showing how the 'free' software isn't so very
> 'free.' That doesn't make the license a 'bad' thing, but shows
> how the license terms tends to help remove degrees of freedom
> (esp for add-on software developers.)
> Simple: using the GPLed compiler as if it is commercial (e.g. carefully
> reading the license and understanding the appropriate limitations) and
> recognize that ownership and/or freedom to capitalize upon significant
> modifications is pre-ordained by the owner (not necessarily author) of
> the codebase. On the other hand, if you start with truly free software
> (or relatively more free), then you can often treat the code modifications
> as if the modifications are yours (which the mods should be.) IMO: Freedom
> should be meant not just for users to do their trade with reasonable freedom,
> and not just distributors to do their trade with reasonable freedom but also
> add-on software developers (whose work is significantly more encumbered
> by GPL rules as opposed to distributors whose (for example) reproduction
> of CDROMs doesn't require giving away their work product IP.) Back
> to facts instead of IMO: Add-on
> software developers do have their add-on work product substantially encumbered
> by the GPL (and the GPL is designed that way.) I am NOT asserting whether or
> not a codebase should/should not be substantially redistribution
> restricted (like GPLed code is), but consistency, honesty
> and following the license terms (Often several pages of text) is important.
> It is also important (IMO) to avoid claiming something is a license of
> 'free' software when the terms aren't really very free for everyone.
Yes, I understand that GPL libraries, for instance, are considerably
restricted by the GPL. One can only hope that such licensors are aware
of and truly wish to impose those conditions. However, this thread is
really about gcc - and gcc does not restrict code generated with it (as
we agree). I did not really intend to rehash the issue w.r.t. libraries
and GPL'd code in general.
> CDROM distributors don't have their work product significantly encumbered
> by the GPL (anymoreso than free software.) End users (who don't incorporate
> GPLed codebases into their product) dont' have their work product significatnly
> encumbered by the GPL. However, add-on software developers (or those
> who contract software developers, but own the work product) do have to offer
> to give away their hard-earned work product upon the distribution of binaries
> of a GPLed codebase. (Specifically, a bona-fide offer of the source
> needs to be provided to those who receive binaries.)
Yes, that is a well understood and deliberate condition of the GPL ...
As in the point above about libraries, if the author does not intend to
apply that condition, they may choose a different license? In the case
of my own work, I do intend to apply those conditions, since I my own
opinions align with the intentions of the GPL.
> Again, I am not making a strong value judgement as to the goodness of
> encumbering the work product of add-on software developers, but calling
> such encumbered codebases as free is certainly not being very honest.
> Sophistry that tries to redefine free also doesn't bode well about the
> (self) honesty of those who do the redefintion.
> > Decline the challenge?
> You didn't really just challenge me, but you were overly rude. Note that
> you are now more clearly challenging with a little less low-class rudness,
> and I have answered (quite accurately.)
> Note that the best way to answer your question (about having the
> restrictive GPL licnese causing troubles in use) is that it forces accomodation
> that wouldn't necessarily be needed (or even very helpful.)
My (unnecessary) challenge was merely whether any of us had been in a
situation where simply using gcc to build a program had encumbered that
program. I guess we're agreed that is not possible.
> Asking the
> question about the GPL keeping people from using the codebase isn't a correct
> question. Note that paying money for a commercial codebase is the similar kind
> of restriction for normal commercial licenses as having to distribute
> (give away) hard-earned intellectual property is for the GPLed codebases.
Yes - once you've bought it, you're usually restricted in what you can
do with it. The GPL is the same except you can always obtain the source
free of charge. Right?
> Again, GPLed compiler codebases can often be treated as if they are commercial
> and proprietary and you won't get in trouble. If you start modifying the
> codebase, then you need to be VERY careful about subsequent distribution
> of the modified GPL codebase. Refer carefully to the multi-page
> and/or multi-paragraph GPL license, understand it then that is similar
> in complexity to multi-page and/or multi-paragraph commercial licenses.
Yes, but no conditions apply to code generated by said modified
compiler? And no conditions apply to the modified compiler source
*unless* you intend to distribute the binary.
> On the other hand, licenses of free software can be much less demanding
> upon those who modify/redistribute the modifications.
> Since most people aren't modifying the GPLed, commercial or free
> (specific case) compilers,
> then you can clearly treat the GPLed compilers as if they are commercial
> and not have troubles. Treating the compiler (with need for potential
> modifications, often paid for at great expense) as if it is truly free can
> cause a large number of complaints from the "GPL police." I am NOT
> judging the desirability of forcing people to distribute (expensive)
> modifications of a codebase (based upon the GPL requirements that
> would then allow distribution of binaries of that modified codebase),
> but mostly warning people NOT to trust GPL advocacy -- esp those who
> react strongly to suggestions that one carefully understand the license.
> I did give a quick (generally valid) short cut to treat GPLed codebases
> like compilers as if they are 'commercial.' The same situations that
> trigger problems for commercial codebases CAN give problems for the
> GPLed codebases.
> MY OWN SUGGESTION is to initially TREAT GPLed codebases as being similar
> to commercial software. This will generally avoid violating the GPL
> encumberances. As you need to distribute your binaries, then more
> more careful understanding of the license is important.
> It is very common to have an overly aggressive GPL advocate to respond
> unfavorably to a strong admonishment to understand the GPL license --
> and don't make assumptions based upon the common usage of the term
As a GPL licensor myself I do make it a point to study the GPL
frequently and refer to it rather than rely on memory. Whether the word
'free' is misapplied or not, the principles and intent of it may always
remain too radical for many people (in particular, BSD license
advocates, who often have their own brand of myopia). Thanks for the