C++11 & The Long-Term Viability Of GCC Is Questioned

[RealNC]

People who dislike the GPL are usually not users. The GPL is extremely permissive to users. That's the whole point of it. And that's what makes this license so great. It gives me permission to do virtually everything I could possibly want as a user. That's much more meaningful compared to the "free beer" approach of the BSD license, which is very restrictive.

People who claim that the BSD license is "permissive" seem to not know what they're talking about. How is it permissive if my ability to freely copy (and/or modify) software that uses BSD licensed code is restricted? That is not permissive at all.

And from a programmer's point of view, why on earth would I put countless hours of work in something and then license it under BSD and let everyone sell proprietary products that use my work while I don't get anything (be it payment or code)? If I instead use the GPL, I know that I can get code, or reserve the right to sell a proprietary license instead so I get payment. Compensation in code or money. With the BSD license, you get neither.

So no, the LGPL is not better for libraries. It's a compromise and shouldn't be used if you can avoid it.

[ryao]

People who dislike the GPL are usually not users. The GPL is extremely permissive to users. That's the whole point of it. And that's what makes this license so great. It gives me permission to do virtually everything I could possibly want as a user. That's much more meaningful compared to the "free beer" approach of the BSD license, which is very restrictive.

People who claim that the BSD license is "permissive" seem to not know what they're talking about. How is it permissive if my ability to freely copy (and/or modify) software that uses BSD licensed code is restricted? That is not permissive at all.

And from a programmer's point of view, why on earth would I put countless hours of work in something and then license it under BSD and let everyone sell proprietary products that use my work while I don't get anything (be it payment or code)? If I instead use the GPL, I know that I can get code, or reserve the right to sell a proprietary license instead so I get payment. Compensation in code or money. With the BSD license, you get neither.

So no, the LGPL is not better for libraries. It's a compromise and shouldn't be used if you can avoid it.

For what it is worth, the GPL is usually only good for users if they know how to compile things themselves or someone is willing to be liable for distribution. I recently built GCC for an obscure platform and someone asked me for binaries. I refused to provide them to him because of the GPL.

[RealNC]

For what it is worth, the GPL is usually only good for users if they know how to compile things themselves or someone is willing to be liable for distribution.

It's also good for users since they can copy the software and give it to others, even if it's commercial. No EULA stuff to worry about.

[frantaylor]

For what it is worth, the GPL is usually only good for users if they know how to compile things themselves or someone is willing to be liable for distribution. I recently built GCC for an obscure platform and someone asked me for binaries. I refused to provide them to him because of the GPL.

So you say that the extra quality that comes from patches pushed upstream, this means nothing to users?

Yes indeed RedHat software generates billions of dollars a year in revenues, and puts themselves into the Dow Jones Industrial average, by selling something that means nothing to their end users.

Maybe you could try to say something that makes less sense, but I think it would be hard.

And by the way, people don't usually BRAG about being butt-wipes. Yes indeed use the work of others for your own ends, and then laugh when asked to be part of the community.

[ryao]

It's also good for users since they can copy the software and give it to others, even if it's commercial. No EULA stuff to worry about.

They cannot do that without being subject to the GPL requirements themselves.

So you say that the extra quality that comes from patches pushed upstream, this means nothing to users?

Yes indeed RedHat software generates billions of dollars a year in revenues, and puts themselves into the Dow Jones Industrial average, by selling something that means nothing to their end users.

Maybe you could try to say something that makes less sense, but I think it would be hard.

And by the way, people don't usually BRAG about being butt-wipes. Yes indeed use the work of others for your own ends, and then laugh when asked to be part of the community.

I have no clue what you mean by "extra quality". I applied no patches to the sources. I just happen to be one of the few people who understands how to build a toolchain using them. Anyway, I am not opening myself to a lawsuit should I be asked for source code at some point in the future and I can no longer produce it. If GCC had been BSD-licensed, I would have happily given that person binaries.

With that said, it would seem that some people will complain no matter what I say. If I say that the GPL prevents developers from wanting to use certain software, they chastise the developers for not embracing GPL software development. If I say that the GPL prevents developers from distributing compiled binaries, they chastize the developers for "not being part of the community". It is ridiculous. Build your own binaries from your own code and leave others out of it. Also, do not publish open source software if you do not want people using your code.

[XorEaxEax]

For what it is worth, the GPL is usually only good for users if they know how to compile things themselves or someone is willing to be liable for distribution. I recently built GCC for an obscure platform and someone asked me for binaries. I refused to provide them to him because of the GPL.

How hard would it be to put your the source tree of your GCC patches onto something like github and direct any binary recipients there, seriously?

Of course, the big draw of GPL from a programmer standpoint is as always that they, as recipients of modified versions of their code are entitled to those modifications in source code form.

They cannot do that without being subject to the GPL requirements themselves.

Which has already been honored upstream, so unless they make modifications they can just point to the upstream source, if they make changes then yes, they must also provide a place for recipients to get the source code modifications they've made, there are tons of places where you can host these modifications online for free.

I have no clue what you mean by "extra quality". I applied no patches to the sources. I just happen to be one of the few people who understands how to build a toolchain using them. Anyway, I am not opening myself to a lawsuit should I be asked for source code at some point in the future and I can no longer produce it.

Unless you've modified the GCC code (including build scripts) then how could you possibly be subject to a 'lawsuit'?

Not that I see why someone would refrain from giving out toolchain build instructions if someone asked for them, but it's not against the law to be a jerk. So if you've made no changes to the original and built your 'binaries' from the exact same upstream source then how could you possibly be in breach of the licence conditions? Just point any binary recipients to http://gcc.gnu.org/gcc-x.x/

[RealNC]

With that said, it would seem that some people will complain no matter what I say.

It's not complaining. It's correcting your claims. You can give a binary of GPL code to someone who asked and no one can sue you for doing so. You seem to be forgetting that GPL means "General Public License". An individual asking you for a binary is not the general public.

[brosis]

GPL prevents to distribute binaries, if the patches contain crappy license like CDDL. The whole point of GPL is to prevent closing the source code down, thats what somebody should nail in his head and shut up for good. You can't redistribute the binaries, because you mixed the license - this prevents binary spread by users, who never care about license and thus secure the market share spread, by making its redistribution much harder. Make note - the proprietary opposite explicitly disallows use, recompilation and reverse engineering altogether, with no source code available.

If you don't mix crap to GPL, you can compile and distribute binaries freely, is this something hard to understand?
Crap like proprietary code, which itself prevents anything, except usage for single licensed user on single platform entity, and which is impossible to open due to all the trade secrets - just look at nvidia and amd right now.

Every time I see some bitchin GPL is not good and is not free - every single freaking time the source of cries is a proprietary developer!

The only licenses that allows proprietary code are BSD-like ones, and thus our ryao, while whining he can't distribute the binary, because he made cocktail of GPL and dirt, may find himself in much more common situation, where he can't distribute the binary and has no access to source code - because the license is proprietary and the software shares 90% code internally from BSD but gives a damn about any "freedoms" ryao whines about.

GPL is good! The only one who land in a trouble with it and start yelling are proprietary developers. Likewise, GPL started especially due to proprietary developers stripping every right from the software.

So if you write closed source zero freedom software and want opensource minions to improve your code, use BSD. If you don't plan to support proprietary crap - use GPL. It will protect you from closing the source down and will make any attempts to mix source impossible, just the way it is designed. It is the only license that protects advertised freedoms and not only advertises them.

[silix]

People who claim that the BSD license is "permissive" seem to not know what they're talking about. How is it permissive if my ability to freely copy (and/or modify) software that uses BSD licensed code is restricted? That is not permissive at all.

the BSD licence - better, culture - stems from the academia, whereas the Berkeley Software Distribution AT&T licensed unix was effectively an academic asset, so putting things in that perspective may give a better understanding

And from a programmer's point of view, why on earth would I put countless hours of work in something and then license it under BSD and let everyone sell proprietary products that use my work while I don't get anything (be it payment or code)?

now, afaik the point of academic research is to be available for everyone for any purpose (no differently from how the knowledge taught at university will be reused in a paid job as well as working pro bono) especially since it's already been paid for with tuition fees, research funds, donations - ironically coming from companies most often than not - knowing this, if you develop sw in the scope of an academic project, most often than not you're not actually seeking additional compensation...
people get conceited over this closed source vs (free/) open source "war" all the time, but if you look at it like an academic (or a pragmatist) would, you may realize that in the grand scheme of things this "war" is entirely irrelevant, and utterly futile in a sense
what matters is only the advancement of humanity as a whole, ie that technology is available to the large public that will make use of it
of course freely available techology is preferable, but in the end who supplies it doesn't really matter to the user, especially when freely available solutions dont exist, or are not convenient
on the other hand, for actual progress to happen, the state of the art has to advance, but "wheel reinvention" aka parallel or subsequent efforts by different developers directed at the same problem (or class of problems) uses resources without advancing the state of the art (and without allowing to consider that problem "solved" and focus on the next) thus diminishing the effetctiveness of those resources and effort
effectiveness that ideally would be maximised if there was exactly ONE (and not more than one) well designed and universally reused worlwide, solution for each problem domain (now, who wouldnt take pride in that one solution being his own?)
in this respect, duplicated work and ideology intermingled with code development, are a worse evil than corporations, and corporation reusing the code is an intended effect, not a "theft"...

If I instead use the GPL, I know that I can get code, or reserve the right to sell a proprietary license instead so I get payment. Compensation in code or money. With the BSD license, you get neither.

there'll always be some third party for whom developing and/or distributing the code is a no-go, so if you use the GPL you are mathematically assured that at least those will have to come up with a different design and / or implementation than yours
which may be exactly what you wanted to achieve, but for people like the above (yes, they exist), it's nothing short of a disgrace
of course in the state of things as of 2013 it's just not possible anymore to have only one "wheel" for every problem - but by going GPL you're not even trying...

[log0]

For what it is worth, the GPL is usually only good for users if they know how to compile things themselves or someone is willing to be liable for distribution. I recently built GCC for an obscure platform and someone asked me for binaries. I refused to provide them to him because of the GPL.

OK. That is just stupid! What exactly has kept you from giving him the binaries and your modified sources (or just the patch with your changes if file size is a concern)? I mean seriously dude!?