Picking a license
in [Python]
|
Prev: 2to3 issues with execfile on python 3.0 on files with encoding
Next: Extract a bordered, skewed rectangle from an image
From: Paul Boddie on 14 May 2010 14:07 On 14 Mai, 19:00, Patrick Maupin <pmau...(a)gmail.com> wrote: > > Would you have agreed had he had said that "MatLab's license doesn't > do much good" and assigned the same sort of meaning to that statement, > namely that the MatLab license prevented enough motivated people from > freely using MatLab in ways that were important to them? Obviously, > it was important enough to enough people that they went and built the > GPLed Octave software, which now emulates MatLab very closely. I don't need to answer your question. It's obvious that the licence doesn't do much good when people seek to create a platform which is genuinely and irrevocably open as a response. That they have done so using the GPL pretty much sinks the previous ridiculous statement about the GPL, too, unless Octave is somehow a bad thing (which is what a certain vendor of proprietary statistics software would have you believe about a certain widely-used statistical analysis tool). Although people can argue that usage of the GPL prevents people from potentially contributing because they would not be able to sell proprietary versions of the software, it has been in no way demonstrated to be universally true that such contributors would contribute more than those who do so because of the copyleft licensing. The creators of Octave are obviously not willing to create (or help create) another system with all the proprietary limitations of MatLab, and why should they be willing? The production of a different "proprietary flavour" of MatLab wouldn't be beneficial to them at all - it might even be detrimental to their project - and might only be marginally beneficial, at best, to existing MatLab customers. [PySide] > Just as there are a lot of proprietary programs that are relatively > useless and *won't* have any GPLed versions written, nobody's going to > waste time rewriting a marginally useful GPLed library just to put a > permissive license on it, either. Unless they really want to release (or encourage the creation of) proprietary software, which is precisely what PySide is all about. (And PyQt is not "marginally useful" - it is a widely-used and widely well-regarded library.) And this apparent overriding need to support proprietary solutions results in different strategies, such as with the Chandler project: because the OSAF wanted to be able to sell proprietary solutions but didn't own all the code, they decided to pick only permissively licensed software for the components of the solution, resulting in a lot of extra effort expended in getting their user interface toolkit up to scratch. You can make your own mind up about whether that was a sensible strategy. Usually, however, most people wanting to write proprietary software cannot be bothered to do the work to replicate an existing GPL- licensed solution (or even to significantly improve permissively licensed solutions). They instead appeal to people to release already- mature permissively licensed software, typically waiting for someone with enough money or manpower to do most of the work for them. Again, this is precisely why PySide appeals to a certain audience. Paul
From: Patrick Maupin on 14 May 2010 14:36 On May 14, 1:07 pm, Paul Boddie <p...(a)boddie.org.uk> wrote: > On 14 Mai, 19:00, Patrick Maupin <pmau...(a)gmail.com> wrote: > > Would you have agreed had he had said that "MatLab's license doesn't > > do much good" and assigned the same sort of meaning to that statement, > > namely that the MatLab license prevented enough motivated people from > > freely using MatLab in ways that were important to them? Obviously, > > it was important enough to enough people that they went and built the > > GPLed Octave software, which now emulates MatLab very closely. > > I don't need to answer your question. It's obvious that the licence > doesn't do much good when people seek to create a platform which is > genuinely and irrevocably open as a response. That they have done so > using the GPL pretty much sinks the previous ridiculous statement > about the GPL, too. That statement was made in the context of why Carl doesn't use GPL- licensed *libraries*. He and I have both explained the difference between libraries and programs multiple times, not that you care. > unless Octave is somehow a bad thing (which is > what a certain vendor of proprietary statistics software would have > you believe about a certain widely-used statistical analysis tool). > Although people can argue that usage of the GPL prevents people from > potentially contributing because they would not be able to sell > proprietary versions of the software, it has been in no way > demonstrated to be universally true that such contributors would > contribute more than those who do so because of the copyleft > licensing. As I have said before, the availability of multiple (but not too many!) licenses is a great thing, because each contributor can decide how he wants to license his creation. Finding the right license to contribute under can only enhance the commons. > The creators of Octave are obviously not willing to create > (or help create) another system with all the proprietary limitations > of MatLab, and why should they be willing? I don't presume to know their motivations, or how the license got chosen. However, once it was under the GPL and there were multiple contributors, it would certainly be difficult to relicense any other way. > The production of a > different "proprietary flavour" of MatLab wouldn't be beneficial to > them at all - it might even be detrimental to their project - and > might only be marginally beneficial, at best, to existing MatLab > customers. I personally can't see any realistic chance of detriment. How could a proprietary clone hope to compete against free software on one side and real matlab on the other side? That's a no-win position, so I wouldn't expect to see any proprietary clones. > [PySide] > > > Just as there are a lot of proprietary programs that are relatively > > useless and *won't* have any GPLed versions written, nobody's going to > > waste time rewriting a marginally useful GPLed library just to put a > > permissive license on it, either. > > Unless they really want to release (or encourage the creation of) > proprietary software. How does recreating something marginally useful encourage proprietary software? That's very confusing. > which is precisely what PySide is all about. No, PySide is about non-GPL software, and is released under a license that even RMS recognizes as "free", and it is certainly not of marginal utility. > (And PyQt is not "marginally useful" - it is a widely-used and widely > well-regarded library.) Well, we agree on that. But I don't know why you're trying to claim I said PyQt was only marginally useful. > And this apparent overriding need to support > proprietary solutions results in different strategies, such as with > the Chandler project: because the OSAF wanted to be able to sell > proprietary solutions but didn't own all the code, they decided to > pick only permissively licensed software for the components of the > solution, resulting in a lot of extra effort expended in getting their > user interface toolkit up to scratch. You can make your own mind up > about whether that was a sensible strategy. Large, high-risk projects are often going to fail and there will always be some decisions that are easy to second-guess, correctly or not. In any case, if the goal was a particular method to get a ROI, it may have been that they wouldn't have been able to do that at all with the GPL, either. > Usually, however, most people wanting to write proprietary software > cannot be bothered to do the work to replicate an existing GPL- > licensed solution (or even to significantly improve permissively > licensed solutions). Usually, most people wanting to write software can't be bothered to do the work. That's nothing new. Yet occasionally people do some work, and some projects make progress. > They instead appeal to people to release already- > mature permissively licensed software, typically waiting for someone > with enough money or manpower to do most of the work for them. Well, personally, I'm still waiting for someone to finish the Hurd, but I've always been led to believe that laziness is a virtue in a programmer. > Again, > this is precisely why PySide appeals to a certain audience. Certainly, the audience for a library is going to be people who both find it interesting and find that it meets their licensing requirements. Nothing new about that. Regards, Pat
From: Paul Boddie on 14 May 2010 14:38 On 14 Mai, 19:15, Patrick Maupin <pmau...(a)gmail.com> wrote: > On May 14, 11:48 am, Paul Boddie <p...(a)boddie.org.uk> wrote: > > Section 3 of GPLv2 (and section 6(d) of GPLv3 reads similarly): "If > > distribution of executable or object code is made by offering access > > to copy from a designated place, then offering equivalent access to > > copy the source code from the same place counts as distribution of the > > source code, even though third parties are not compelled to copy the > > source along with the object code." > > > And here's that FAQ entry which clarifies the intent: > > >http://www.gnu.org/licenses/gpl-faq.html#DistributeWithSourceOnInternet [...] > http://www.gnu.org/licenses/gpl-faq.html#UnchangedJustBinary We're all aware of the obligation to provide source code. You've spent the last few days complaining about it. > > Like I said, if you really have a problem with Ubuntu shipping CDs and > > exposing others to copyright infringement litigation. > > So, deliberately or not, you're trying to change the discussion > again. I *never* discussed Ubuntu shipping a physical CD, and never > intimated that that was a problem. My discussion was *always* about > an individual *downloading* an ISO and *burning* a CD himself, then > *distributing* the CD to someone else. I am not changing the discussion at all. You are describing a situation where someone gets the binaries but not the sources, but according to the licence they should get both of those things (ignoring written offers and the like), and this does apply to Ubuntu since precisely this act of distribution (to use the older term) is performed by them. That you then pass on the binaries without the sources is an equivalent situation, ignoring for the moment that you do not yourself have the sources either. So, what are you supposed to do when the recipient "calls" you on the lack of sources? (And, yes, clearly the FSF anticipates that not everyone will request the sources because it is written in that very excerpt I provide above.) If the recipient is strict about exact compliance, you will have to provide the sources on CD to them. And this makes sense: if they can only make use of the binaries if provided on CD (and not, say, on an FTP site because they don't have an Internet connection, for example), then they will need to receive the sources in the same manner. Of course, the recipient may only demand certain sources, not wishing to avail themself of the sources for all copyleft-licensed packages in the binary distribution. Now we return to the matter of getting the Ubuntu sources. If you ordered a CD from Ubuntu via their ShipIt service, it is at this point that you can demand a CD of corresponding sources. If they cannot provide one, then obviously it poses a problem for your compliance (and theirs, and you should see once again why Ubuntu's activities do matter), but naturally Ubuntu provide parallel binary and source repositories for all their packages. So, even if they were found not to be in compliance according to the strictest interpretation of the licence, it is technically possible for you to acquire the corresponding sources and make them available to the person who was given the CD. If you downloaded an ISO file, Ubuntu could (and do) obviously provide source packages from the same location: their Web site and various mirrors. Really, if at this point you think I'm playing games with you, then you really need to stop taking score and formulate the exact problem you have with the distribution of Ubuntu-style media, because I'm starting to think that the only real problem here is the one you have with people using copyleft-style licences for their works. Since we've had to hear about that over several days, I don't think that articulating that particular problem once again really brings anything more to the discussion. Paul
From: Patrick Maupin on 14 May 2010 15:14 On May 14, 1:38 pm, Paul Boddie <p...(a)boddie.org.uk> wrote: > On 14 Mai, 19:15, Patrick Maupin <pmau...(a)gmail.com> wrote: > > > On May 14, 11:48 am, Paul Boddie <p...(a)boddie.org.uk> wrote: > > > Section 3 of GPLv2 (and section 6(d) of GPLv3 reads similarly): "If > > > distribution of executable or object code is made by offering access > > > to copy from a designated place, then offering equivalent access to > > > copy the source code from the same place counts as distribution of the > > > source code, even though third parties are not compelled to copy the > > > source along with the object code." > > > > And here's that FAQ entry which clarifies the intent: > > > >http://www.gnu.org/licenses/gpl-faq.html#DistributeWithSourceOnInternet > > [...] > > >http://www.gnu.org/licenses/gpl-faq.html#UnchangedJustBinary > > We're all aware of the obligation to provide source code. You've spent > the last few days complaining about it. > > > > Like I said, if you really have a problem with Ubuntu shipping CDs and > > > exposing others to copyright infringement litigation. > > > So, deliberately or not, you're trying to change the discussion > > again. I *never* discussed Ubuntu shipping a physical CD, and never > > intimated that that was a problem. My discussion was *always* about > > an individual *downloading* an ISO and *burning* a CD himself, then > > *distributing* the CD to someone else. > > I am not changing the discussion at all. You are describing a > situation where someone gets the binaries but not the sources, but > according to the licence they should get both of those things > (ignoring written offers and the like), and this does apply to Ubuntu > since precisely this act of distribution (to use the older term) is > performed by them. That you then pass on the binaries without the > sources is an equivalent situation, ignoring for the moment that you > do not yourself have the sources either. If Joe downloads and burns a CD for his friend, he may not have the sources and may not have any intention of getting them, and probably didn't provide a "written offer." What you're "ignoring for the moment" is my whole point, that unlike Ubuntu, Joe is now in violation of the GPL license, because he provided neither a written offer nor source on CD, nor his own download site. > So, what are you supposed to do when the recipient "calls" you on the > lack of sources? There is possibly no "calling". Since no source and no written offer was delivered, Joe's friend may not know about the issue. Even Joe himself just saw it was "free software" and didn't read the fine print, so he may not have a clue how to get the source. >(And, yes, clearly the FSF anticipates that not > everyone will request the sources because it is written in that very > excerpt I provide above.) Which Joe doesn't know about and didn't adhere to in any case. > If the recipient is strict about exact > compliance, you will have to provide the sources on CD to them. That could be a year later, and Joe, who doesn't really even know anything about source, is really going to have a hard time figuring out exactly which sources went into the CD he downloaded that long ago. > And > this makes sense: if they can only make use of the binaries if > provided on CD (and not, say, on an FTP site because they don't have > an Internet connection, for example), then they will need to receive > the sources in the same manner. To an extent it makes sense. That's why I explained that I thought it would be nice of Ubuntu to put a warning to Joe on their site explaining the consequences of helping his friend out. Of course, since the warning would only serve to decrease object downloads, and since Joe's friend doesn't really want the source anyway, there is no real point. That doesn't alter the fact that Joe is immediately in violation of the GPL once he delivers the CD to his friend without the written offer. > Of course, the recipient may only > demand certain sources, not wishing to avail themself of the sources > for all copyleft-licensed packages in the binary distribution. [ Stuff about ShipIt snipped because I was never discussing that.] > Really, if at this point you think I'm playing games with you. I don't know what to think about that. Even after I've explicitly said multiple times I'm not discussing when Ubuntu ships a CD, you still felt compelled to include a big paragraph about ShipIt. Is it to confuse? Or because have OCD? I don't really know. > then > you really need to stop taking score and formulate the exact problem > you have with the distribution of Ubuntu-style media, I explained it fully multiple times. > because I'm > starting to think that the only real problem here is the one you have > with people using copyleft-style licences for their works. Well, as I have tried to explain, there are tradeoffs with any license, including the GPL. With the GPL, you can easily adhere to the letter of the license by shipping source with object. But sometimes the source is so huge, people take shortcuts to get around that. The license allows this, via a requirement to provide source later, that could actually fall on people who really don't understand that that is required of them (perhaps not very often in practice, but certainly in theory). > Since we've > had to hear about that over several days, I don't think that > articulating that particular problem once again really brings anything > more to the discussion. Probably not, but you've still never addressed how easy it is for a complete neophyte to go to, e.g. the Ubuntu site, be told do a download, try it out, get excited, burn CDs for his friends, and then be in violation of the GPL. As I have said, I don't view this as a *practical* issue, but it is an example of how those in the GPL community turn a blind eye to innocent infringement, just like Microsoft. After all, the Ubuntu download page says "When the CD is ready, simply put it in your CD drive, restart your computer and follow the instructions that will appear on your screen. Don't forget that you can create more copies and pass the CD to as many people as you like." without mentioning *anything* about source code. Regards, Pat
From: Paul Boddie on 14 May 2010 15:26
On 14 Mai, 20:36, Patrick Maupin <pmau...(a)gmail.com> wrote: > > That statement was made in the context of why Carl doesn't use GPL- > licensed *libraries*. He and I have both explained the difference > between libraries and programs multiple times, not that you care. Saying that GPL-licensed applications are acceptable is a minor concession to the use of copyleft licensing if one advocates permissive licensing for all things which are not perceived to be finished products: things that one isn't looking to re-use somehow. Saying that one likes Octave and that it uses the GPL, too, is really damning it with faint praise if one were then to say that its parts should be permissively licensed so that one can incorporate its functionality into something else. No, I don't care if you have a problem with GPL-licensed libraries because it is, as we have established repeatedly, your problem not mine. [...] > > The production of a > > different "proprietary flavour" of MatLab wouldn't be beneficial to > > them at all - it might even be detrimental to their project - and > > might only be marginally beneficial, at best, to existing MatLab > > customers. > > I personally can't see any realistic chance of detriment. How could a > proprietary clone hope to compete against free software on one side > and real matlab on the other side? That's a no-win position, so I > wouldn't expect to see any proprietary clones. Well, only permissively licensed software would encourage such clones. At that point, there are incentives for people to develop functionality for proprietary deployment instead of for the upstream project. [PySide and proprietary software] > No, PySide is about non-GPL software, and is released under a license > that even RMS recognizes as "free", and it is certainly not of > marginal utility. No, PySide is about permitting the development of proprietary applications by providing a solution to the all-important "ISVs" which lets them develop and deploy proprietary software. Do you really think a platform vendor whose "ISVs" routinely ship proprietary software on their platform and on other platforms, and who will demand the ability to continue to do so, now expects all these "ISVs" to provide their applications under the modified BSD licence? Sure, other developers can use the software - even people releasing GPL-licensed software - but that is highly unlikely to be the primary business motivation. If you think the mobile telephony vendors are a bunch of fluffy bunny rabbits playing with each other in sugary meadows of niceness, I don't want to be present when someone directly and finally disabuses you of this belief. It's all about people selling stuff to "consumers" over and over again, preferably with the "consumers" rarely if ever being able to opt-out and do things their own way. > > (And PyQt is not "marginally useful" - it is a widely-used and widely > > well-regarded library.) > > Well, we agree on that. But I don't know why you're trying to claim I > said PyQt was only marginally useful. Because you followed on from writing about PyQt by introducing the topic of "marginally useful" libraries, thus giving the impression that you regarded PyQt as "marginally useful". Paul |