Brian Thomas Sniffen wrote: > Josh Triplett <josh.trip@verizon.net> writes: >>Brian Thomas Sniffen wrote: >>>Josh Triplett <josh.trip@verizon.net> writes: >>> >>>>That's only the case if you consider the right to take the work >>>>proprietary useful, and helpful to Free Software. >>> >>>Or helpful to users. >> >>Users who want to write proprietary software can figure out for >>themselves which software allows them to do so, and which does not. I >>don't think we should be in the business of promoting licenses that help >>people write proprietary software. > > That position is rather more extreme than Debian's historical > standard. "Our priorities are our users and free software," right? > If it wasn't convenient to use Debian for producing proprietary data, > it would see very little use. Imagine a license which compelled me to > publish all data processed with a program and later distributed -- > that is, take the model you want to license compilers under and extend > it to word processing. How about gnucash and its remote-payment > capability? > > An important part of freedom is each person's ability to choose or > reject it for himself. By all means, exhort others to copyleft their > data and programs. I'll be right there with you. But compulsion to > copyleft is repugnant to me, and non-free. Of course. I was not suggesting that Debian should actively hinder those who want to write proprietary software (beyond using copylefts for Debian's own software), just that we don't need to go out of our way to single out which software can and cannot be linked into proprietary software for the benefit of those who want to make that software proprietary, and that Debian should not favor a license because it allows you to take a work proprietary. However, I should have clarified: software that says "you can only _use_ this to make Free Software" is non-free, ironically, since that is a use restriction. I should have said that users who want to _take software proprietary_ can figure out for themselves which software allows that and which does not. >>>>I consider it to be neither. In my case, I would have absolutely no >>>>interest in taking the software proprietary, so that "right" would >>>>be worthless to me, and I would prefer that as few people as >>>>possible have the right to take my code proprietary. Ideally none, >>>>of course. >>> >>>I think there's some fuzziness here about "take code proprietary". >>>Do we all mean distribution under a non-Free license, or without source? >> >>Either would be proprietary. By proprietary, I mean "not Free >>Software". There are two disjoint sets covering all works: proprietary >>and Free. > > Sorry, I meant that as one option, in contrast to others such as "I > have a copy and he has a copy, and we both have freedom, but you don't > have a copy" on one hand, or "I have a copy and freedom, and it's > available to the world on this web server" on the other. Ah, I see. Good point. Yes, I consider software distributed privately to still be Free Software as long as all those with a copy have freedom over that copy. >>>>>some people don't think it's non-free -- "If I can make the modifications >>>>>guaranteed by the DFSG, what's the harm?", but one of the real benefits of >>>>>Free Software is that no member of the community has an inherent advantage >>>>>over anyone else -- a "free market" ideal. >>>> >>>>I consider the benifit of Free Software to be that everyone has all the >>>>essential freedoms over the software. If some people have non-essential >>>>freedoms, such as the "right" to take the software proprietary, then >>>>that is irrelevant to me. >>> >>>So if I compel you to allow *your* software to be taken proprietary, >>>that's free? >> >>My own software, unrelated to yours? No, and you would have no legal >>ability to do so. My software which is a derivative of yours? That >>would be acceptable, just obnoxious. > > Why is that acceptable? I understand that copyright law allows it, > but it seems very clear to me that it's a payment: I will license my > software this way, and in exchange you will give me a license that > way. Consider it separate from the QPL's non-free restrictions. I am not required to distribute the software to you at all. If I do, you happen to get permission to take the software proprietary (in the case of the QPL, only if you also continue distributing under the QPL); however, if you licensed it under a license that forced me to give those permissions to everyone (such as a hypothetical free version of the BSDPL), I would also be compelled to give you that same permission, along with everyone else. Also see the DFSG FAQ (http://people.debian.org/~bap/dfsg-faq.html) point 12e. It isn't non-free to say "Free Software developers can do this" or "people named Bob can do that" or "the initial developer can do such-and-such"; as long as everyone can still exercise all of their freedoms, it's acceptable if some people have additional permissions. > For example, I have a big library of code. I want to embed it in a > compiler, to produce proof-carrying code. I wrote that library on my > own. But if I break it up into little bits, and embed that throughout > a compiler -- say the OCaml compiler -- and distribute it as > patches... why is it free for INRIA to tell me how to license that? I > had it before hand, I wrote it without reference to their code, and > it's useful in its own right. > > When combined with theirs, sure, the whole work has to be distributed > only under terms that both they and I find acceptable. But the QPL > asks more than that: it says that my part, distributed alone as > patches, must be under a license they find acceptable. If your work isn't a derivative of OCaml, and you aren't distributing it as part of OCaml, you can license it however you like. If you distribute it as part of OCaml, or you distribute it separately but it is a derivative of OCaml, then they can certainly tell you what rights you must grant for users of your work. >>>1: Because the "initial author" hasn't changes and your work is only >>> distributed as patches, you can't get the same sort of QPL the >>> original author used, only an inferior-QPL in which you don't get the >>> free license and to demand linked works. >> >>That's an interesting point. By that argument, the QPL blatantly fails >>DFSG3, because you can't distribute your modifications under the same >>license as the original. >> >>I'm more inclined to believe, though, that if you really wanted to, you >>could license your modifications (meaning the patch you distribute) >>under the QPL, and people downstream from you would have to distribute >>patches to your patch, provide changes on request, etc. > > That requires ignoring the phrase "initial developer". I don't think > they'd have repeatedly said "initial" if they didn't really mean it. For the purposes of the original work, they are the initial developer. For your changes, you are the initial developer. >>>And we can substitute any permissive license for "BSD" and any >>>copyleft for GPL up there. The QPL shares a characteristic with the >>>BSDPL: that it forbids you to copyleft *your* work. This seems very >>>different from an author's decision not to copyleft his work. >> >>No, I don't think so. The QPL is a copyleft, which happens to also >>include a more permissive license to upstream. It's also non-free for >>other reasons, but I don't think this is one of them. > > It doesn't just include a more permissive license; that would be a > reasonable phrasing if I chose to distribute my modifications to some > GPL'd software as GPL to the world and BSD to upstream. You could say > that I would include a more permissive license to upstream. When I > issued that license, it would be free. But this *compels* a more > permissive license to upstream. That's a nitpick of how I worded the point. How about "The QPL is a copyleft, which also grants the initial developer additional permissions others don't have." Again, I'm not arguing that the QPL is Free, just that that particular clause is OK. >>>Even the GPL allows me to distribute *my* work separately from the >>>original, and under the BSD license. I can even distribute them together, >>>with mine under the BSD license and the combination and the original >>>under the GPL. The QPL is doing something relatively new and >>>different, in that it compels me to license my work in a particular >>>way, not just the original or a derivation. >> >>Your work that you are referring to _is_ a derivation, is it not? > > No. It's my components of a work which, when combined, is a > derivation. But since I have to ship it as patches anyway, it's very > easy to show what's mine and what's theirs. Work that is entirely written by you can still be a derivative of another work. For example, if you write a program that uses GNU Readline, that program is a derivative of GNU Readline, even if you don't actually distribute GNU Readline with your program. - Josh Triplett
Attachment:
signature.asc
Description: OpenPGP digital signature