Re: OSD modification regarding what license can require of user
On Mon, 2002-03-18 at 18:01, phil hunt wrote: This ties it to a specific technology. For all anyone knows, no-one will be using http in 109 years time. Once HTTP goes away (which will probably be 109 years) change the protocol in the license. The point is that we want to enforce distribution of source not provide a potential pandoras box. Let's examine another scenerio. What if I write some sort of web service and I have my Obtain Source button require that you register your name in my SPAM mailing database. Am I allowed to do that? What language will you introduce to defend against that? The problem is that this approach allows software authors to add sections of immutable code as long as they are source distribution facilities. I think it is far more simple to decouple distribution enforcement from the actual features of the program. -- _ Ean Schuessler [EMAIL PROTECTED] Brainfood, Inc. http://www.brainfood.com -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
On Tuesday 19 March 2002 3:48 pm, Ean Schuessler wrote: On Mon, 2002-03-18 at 18:01, phil hunt wrote: This ties it to a specific technology. For all anyone knows, no-one will be using http in 109 years time. Once HTTP goes away (which will probably be 109 years) OK, I meant 10 years. Slip of the finger. change the protocol in the license. The point is that we want to enforce distribution of source Then have some words to the effect bthat the system uses usual and convenient technological methods to distribute it. -- Philip Hunt [EMAIL PROTECTED] I would guess that he really believes whatever is politically advantageous for him to believe. -- Alison Brooks, referring to Michael Portillo, on soc.history.what-if -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
RE: OSD modification regarding what license can require of user
On Fri, 2002-03-15 at 18:53, Lawrence E. Rosen wrote: My concern is only with the interaction of that requirement with Bruce Perens' proposed OSD change. How are we to decide, a priori, whether a license condition imposed upon a licensee is reasonable or burdensome? Is it reasonable or burdensome, as you proposed, to require users who are ASPs to release their versions with a download server source button prominently located on pages every user sees? All pages? What are permissible requirements without exceeding the bounds of good taste? I agree. Licenses should address legal events such as distribution. When they impose structural requirements on the code or the interface is when they can become awkward or simply absurd. The RPC point is certainly valid. What if someone wants to use the Slashdot source and to provide a service that distributes RSS news feeds as the result of SOAP requests? What if you want to aggregate several GPL web services into a new single service. Does it have multiple view source buttons? What if you want to distribute that service to a device with limited screen real-estate, like a cell phone? It seems too easy to find problems with this solution and the real world is a lot more creative than I am. What if you simply added a requirement that: http://[service host name]:80/gnu-sources Must always either supply the sources or a redirect to the sources? This rule could even apply for internal distribution (ie. services only available to AOL users). That would seem to take care of the problem without placing potentially unfulfillable constraints on the user interface. -- _ Ean Schuessler [EMAIL PROTECTED] Brainfood, Inc. http://www.brainfood.com -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
RE: OSD modification regarding what license can require of user
On 18 Mar 2002, Ean Schuessler wrote: What if you simply added a requirement that: http://[service host name]:80/gnu-sources Must always either supply the sources or a redirect to the sources? This rule could even apply for internal distribution (ie. services only available to AOL users). That would seem to take care of the problem without placing potentially unfulfillable constraints on the user interface. It assumes http. Imagine if someone had put a requirement in ten years ago which mandated that the sources be available by a Gopher link. -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
On Monday 18 March 2002 5:14 pm, Ean Schuessler wrote: What if you simply added a requirement that: http://[service host name]:80/gnu-sources Must always either supply the sources or a redirect to the sources? This ties it to a specific technology. For all anyone knows, no-one will be using http in 109 years time. -- Philip Hunt [EMAIL PROTECTED] I would guess that he really believes whatever is politically advantageous for him to believe. -- Alison Brooks, referring to Michael Portillo, on soc.history.what-if -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
RE: OSD modification regarding what license can require of user
When using the expression derivative work you are including publication/distribution: That is: you read my phrase as if I was describing a published-distributed-derivative work. No, I understand you correctly, but you are simply in error. The U.S. copyright law speaks of preparing derivative works as a right reserved to the author. It does not speak of distributing or publishing them. Among references given by other members of this list there is also an important distinction between copyright and ownership of material http://www4.law.cornell.edu/uscode/17/202.html The Title 17 has the same squeletton that others european IP laws The word Copyright is used where Moral rights appears in European corpus. This is the Berna and following Treaties, in fact. All the texts are at : http://clea.wipo.int/lpbin/lpext.dll?f=file[ibrowse-j.htm] (Look under English/WIPO folder) (Note, i understand the classic difference between US/Euro laws not on copyright/moral rights (Art.6), but on the Droit de suite and the inalienability concept of european authors (Art. 14ter of Berna) European authors can forbid the use of their work, even after the cession of their rights. But these are 2 different, albeit complementary, concepts. While the transformed work is not made publicly available, an author cannot forbid transformation. This would be an absurd. Absurd or not, it is the law. Uh ! I believe that laws are not absurd, or they would not be Laws. Imagine the time before Berne Convention. We are in 1885. An US translator take Jules Verne books, publish them under his name and got the US Laws protection. This is legal, because the author protection come from national law. The rights of second author are equally protected. But in 1886, when oversea commercial communications are yet usual, this state of the art of copyright does not appears very moral for the authors (nor editor companies, of course) :)) When the Convention arises, the handshake, and the squeletton of ideas is as follow: a- Moral rights (copyright) are universal and attached to individuals. A good example of copyright is that the name of author must appears on the published work, and before the name of translator. In OS licences, this is for example the prerequisite that the copyright notice must appears on the protected software. Another copyrigh (Moral right) is that translators must ask for author's permission before to _publish_ their translation b- Given the author persmission, then US translator will be granted the same rights that any authors on his own country. c- National Laws preserve the rights of their citizen to translate the Jules Verne books to know their contents, as long as these versions are not published nor used for commercial distribution. In all cases, distribution for State Libraries, for example, are free of commercial/distribution rights Note that the precedence order of the above is : (c) (b) (a) :) Now, consider that translation is the mother of all derivations, and you got my point about software For this same reason, he cannot forbid the _translation_ of his work. (Can you believe seriously that States will adhere to a Convention that allows authors to forbid translation of works ? I do seriously believe it; in fact, I know it is so. Of course, you don't : Based on the (c) provision above, Berna Conventions allows States to distribute works without fees nor royalties in their own countries, as long as their belong to countries under economical developpement. Author's rights come after public interest of States. If you consider that Software is under the Berna protection scheme, there is no reason to consider differently the software industry (I understand that politically, or economically, there are some big questions, here, but we have not to consider political nor economical criteria, for now :)) Anyway, the question is to determine if software may fully adhere to the artistic IP protection scheme, as defined by Berna Convention. The Berna protection implies some limitations for proprietary software (and also some warnings to open source software). It's funny to see how these companies, after their exploit and almost destruction of the Berna palace, try now to send Berna to the trash, seconded by Medical, Music and Video Industry, and move now to patents ! And please note that these companies are working directly at WIPO level. Not at US Laws level. In my opinion, any Open Source licenses should enforce the idea that software is protected under Berna and work this way. If you do not comply with Berna Convention, you are giving arguments to propietary software constructors to adopt the patent strategy And the license protection scheme will be dead. Patents are not copyrights; the rules are entirely different. In particular, there is AFAIK no international patent law at all. Not yet. And this is the problem they
Re: OSD modification regarding what license can require of user
David Johnson wrote: I was not judging any type of application. Instead I was prognosticating from my own particular perspective. No offense was intended and I hope you don't take it as such. I do foresee a need for web services and other forms of centralized processing. But I still doubt that they will become the dominant mode of application computing. We see eye to eye on that; I don't think it'll be the norm for a long time, if ever. Still, for lots of people, GPL software isn't the norm either, and still we spend time trying to hammer down a good license. Likewise, even if the group of applications that uses strict fat server/thin client type of deployment is minute, these applications can (and do, from my perspective) serve an interesting niche; to make it interesting/feasable for people to release such applications this too deserves a good license. I get the impression that there are one or two people out there (not you) who are in a state of urgency over this issue because they see an immanent return of the dumb/thin client. I don't like decisions to be made in panic mode, so I was trying to explain in my own not-to-brilliant way why I didn't think it's going to happen. OK, I see. Two discussions are crossing over into each other here. I don't think that the bulk of applications is going to move to ASP-type software. So on this we agree. Even so, as said, an interesting portion _will_ (yes, like mine), and we need to think about this. Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Russell Nelson wrote: It would be wrong to require publication of modified versions that are used privately, but inviting the public to use a server is not private use. I'm not sure that the GPL-using community is going to agree with you on this. For example, if someone decides to distribute Linux under the GPLv3, and someone else runs it on a server, are they bound by the GPLv3 to become a Linux distributor? I haven't seen it, so I'm just speculating. Only if the Linux kernel (are we talking kernel or distro here?) displays the notice in the first place, I think. Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Russell Nelson scripsit: How do you know that such an intention was there? Perhaps the program was designed to link with another, compatible non-GPL DLL, and mere compatibility allows it to link with the GPL DLL? I really don't know what a judge would say, and I expect that you don't know either. The devil himself knoweth not the mind of man. Legal intent is always judged by objective evidence. -- John Cowan [EMAIL PROTECTED] http://www.reutershealth.com I amar prestar aen, han mathon ne nen,http://www.ccil.org/~cowan han mathon ne chae, a han noston ne 'wilith. --Galadriel, _LOTR:FOTR_ -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
I.R.Maturana scripsit: When using the expression derivative work you are including publication/distribution: That is: you read my phrase as if I was describing a published-distributed-derivative work. No, I understand you correctly, but you are simply in error. The U.S. copyright law speaks of preparing derivative works as a right reserved to the author. It does not speak of distributing or publishing them. While the transformed work is not made publicly available, an author cannot forbid transformation. This would be an absurd. Absurd or not, it is the law. For this same reason, he cannot forbid the _translation_ of his work. (Can you believe seriously that States will adhere to a Convention that allows authors to forbid translation of works ? I do seriously believe it; in fact, I know it is so. BTW, the actual attempt to adopt Software patents arise because companies have discovered that they were unable to forbid the reverse engineering. They cannot sue anybody for this. Patents are not copyrights; the rules are entirely different. In particular, there is AFAIK no international patent law at all. I suppose you are a lawyer. I am only studying these laws since 1998. I am not a lawyer. -- John Cowan [EMAIL PROTECTED] http://www.reutershealth.com I amar prestar aen, han mathon ne nen,http://www.ccil.org/~cowan han mathon ne chae, a han noston ne 'wilith. --Galadriel, _LOTR:FOTR_ -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Russell Nelson wrote: Has A violated the license? That would seem weird, because it pertains clear written instructions (even if those instructions are in fact a shell script). And since B didn't agree to anything, he isn't bound by the 'click-wrap' license. The dragon attacks: Ask your lawyer to include terms in the license that prohibit disclosure of the unpacking method. The license will also have to prohibit distribution of the enclosed tarball. I don't have a lwyer, and frankly, I'm loathe to hire one just so I can release my sources to the public. If an existing license covers my needs, excellent, but if not, I simply won't be releasing it. Not that the world would collapse without access to my software, but I still think it's a shame. Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
David Johnson wrote: My point was that distribution will not become irrelevant just because of the Web Services hype. In the reasonably foreseeable future, most programs will continue to execute locally because of bandwidth constraints. It's more than bandwidth contraints. A future where all applications run on remote servers and all the clients are thin is foreseeable, but not likely. We had that once, back in the days when RMS decided not to place restrictions on usage but only on distribution. We've moved beyond that. We now realize that it is more efficient to have a thousand CPUs process a thousand applications than to have a single centralized CPU process those same thousand applications. Personally I don't think this latest big push towards thin clients is going to be any more successful than the earlier pushes. Some of it will of course happen, but I don't expect it to be the common mode of executing applications. I do foresee distributed processing. But this involves distribution, so no problem. But centralized processing has come and gone. From your point of view, maybe. _I_ do have a web-service type application that I wish to release as open source, and this application will typically be used _exactly_ in the way you say that 'has come and gone'. Whether or not you feel that this is or isn't a good way to deploy applications is entirely irrelevant to me. I have an application to release; if I can't release it under an open source license because these types of apps are deemed irrelevant by the OSI then so be it. Closed source it is. I still have hopes for the amended GPL or a variant of the APSL or W3C license, though. Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Richard Stallman wrote: A simple example: it is totally trivial on Windows to build a 'service' from a DLL, exposing its entire interface. This would be running as a separate executable, but would look like a regular library to any windows program. The FSF's position is that the GPL applies to any programs which are designed to link with that DLL, that this is legally equivalent to statically linking them. OK: it is equally simple to expose the same functionality under a totally different API. For any GPL dll I find, it would take me all of 30-60 minutes to generate a 'server' that calls the DLL for me; my apps connecting to this 'proxy' would talk to this proxy API, which is drop-dead simple under windows. To what extent can these apps be said to be designed to link to this DLL? And what if I build an alternate implementation of this proxy API which does a faux implementation of this proxy API (or just a really crappy one slapped together during lunch)? I could then claim that my program is designed to link to 'whichever windows service implementing this API', and what do you know, there's 'several' of them available, _both_ GPLed. And my app using them doesn't need to be. Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
A simple example: it is totally trivial on Windows to build a 'service' from a DLL, exposing its entire interface. This would be running as a separate executable, but would look like a regular library to any windows program. The FSF's position is that the GPL applies to any programs which are designed to link with that DLL, that this is legally equivalent to statically linking them. -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
The reason we've decided that this ASP requirement is legitimate is that it is a matter of requiring making the modified source code available in a case of public use. It extends existing GPL requirements coherently to a new scenario of usage. It would be wrong to require publication of modified versions that are used privately, but inviting the public to use a server is not private use. -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
RE: OSD modification regarding what license can require of user
I.R.Maturana scripsit: This is a good point, but it need to be developed. Let me add that once permission is given, Copyright Law can give to second author of derivative work the SAME rights than the first author, on their respective versions. J.Cowan Correct. This is a point that is often not understood. Rights of licensees on their respective work will be the SAME, _without prejudice_ of the rights of the author, AND ONLY IF author authorizes the derivation. IRM This solution [The OS mechanism] fails (or will fail) because the rights on derivative works are not under control of first author. Everyone can derive a work, as long as it do not try to distribute his version. J.Cowan No, this contradicts what you say above and is not true under either U.S. law or the Berne convention. The *exclusive* right to create derivative works is one of the copyright owner's rights. For example, you cannot translate a novel from French to English without the consent of the copyright owner (initially the author) of the French version. Nor can you make a book into a film without the agreement of the copyright author. IRM You are right. I am not using correctly the English words (I know). Help me to explain this little, but very important difference: When using the expression derivative work you are including publication/distribution: That is: you read my phrase as if I was describing a published-distributed-derivative work. This way, an author _may_ forbid derivation. But I used the expression derivation with a more restricted meaning: I did not included the publication/distribution step. Let us choice the word transformed, rather than derived. This will clarify the point. While the transformed work is not made publicly available, an author cannot forbid transformation. This would be an absurd. If an author wants seriously to forbid people to transform his work, the only solution is to do not publish nor distribute his work. The point is that once the work is published, once it is made publicly accessible, this public work determines rights for authors AND users. These rights are slightly different, but are _equally_ protected by laws. For example, an author cannot forbid an user to make reverse engineering, look inside the source, modify and even create locally, on his home computer, a new version of the source work. For this same reason, he cannot forbid the _translation_ of his work. (Can you believe seriously that States will adhere to a Convention that allows authors to forbid translation of works ? :-D)) Hope that now, my last message can be better understood. I was trying to show the difference between the distribution rights and the right of users to investigate in any way they like, to assimilate and to transform the original work. This last right is the _same_ for authors and users. IP Laws are not a religion, an author is not the Pope. BTW, the actual attempt to adopt Software patents arise because companies have discovered that they were unable to forbid the reverse engineering. They cannot sue anybody for this. There is a crime, only if the folk try to publish his research. There is a fundamental difference between the citizen's right to open, look inside (and of course, translate) a work without permission of the author, and the author's right to forbid the distribution/publication of the possible derivatives. When Software companies understood this boundary of the IP law, and discovered that IP law would not let them manage the information and ideas as other companies can manage goods, they changed their strategy. The solution is simply to change the protection scheme for software and they are consequently trying to place the software under the industrial patent protection scheme. Anyway, their first attempt to modify IP Laws at the end of 90'th introduced serious limitations to the rights of software authors. Today, software authors rights are really light, compared to artists: number of years before the work fall under public domain, protection Law for companies against developpers, etc. IRM Rights for distribution and for derivation do not come from the same source. J.COWAN They do: the Berne Convention and the various national copyright acts. Berna Convention recognizes 2 agents: Authors, who own their respectives versions, and States from which the authors are also citizens (Les auteurs, du fait de leur creation respective, et les Etats dont ils sont aussi des citoyens/Autores, por el hecho de su creacion respectiva, y los Estados, de los que tambien son ciudadanos) Rights on distribution are rights attached to the author. This is nothing else that the well-know right of property, attached to intellectual goods. On the other side, the right to learn, free speech, open access to sources and ideas, are citizen rights protected by States. These rights define in fact the boundaries of the public space in each country Laws. Strictly
RE: OSD modification regarding what license can require of user
Hi Eben, I appreciate the attempt of the FSF to bring ASP uses under the publish-the-source requirements of the GPL. It is certainly within the rights of a licensor to impose those conditions upon licensees and downstream distributors of modified code. My concern is only with the interaction of that requirement with Bruce Perens' proposed OSD change. How are we to decide, a priori, whether a license condition imposed upon a licensee is reasonable or burdensome? Is it reasonable or burdensome, as you proposed, to require users who are ASPs to release their versions with a download server source button prominently located on pages every user sees? All pages? What are permissible requirements without exceeding the bounds of good taste? I look forward to reading your actual license text. /Larry Rosen -Original Message- From: Eben Moglen [mailto:[EMAIL PROTECTED]] Sent: Thursday, March 14, 2002 6:12 AM To: Emiliano Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: Re: OSD modification regarding what license can require of user We are helping a third party to incorporate our proposed solution to the ASP problem in its own modified GPL for release very shortly (probably within days). Legal work is complete and only some public information documents are not yet final. Use in an FSF-approved third-party license will be followed by inclusion of the term in a draft of GPLv3 that we hope to release for public comment later this year. Our change is not a state secret. Richard has described how it works. If you want to make a new web application's code fully free in ASP use, you release your first version with a download server source button prominently located on pages every user sees. The new license provision extends GPLv2's rule that you cannot remove the copyright notice display from an interactive program to say that if you modify the code you cannot remove the download server source button and functionality. In this way, license terms constrain only modification, in a fashion completely compliant with FSD and OSD. Existing applications are unaffected. New applications and new versions of existing applications can be written so that someone who offers application services must also distribute the conforming source code to all users. Because we have a partner here, I don't want to prerelease the text of its license. If you plan a release within a week, I will be happy to give you my earlier drafts of the language, and help you to make a modified GPL of your own. The whole matter will be completely public so soon, however, that you will almost certainly prefer to wait for announcement. On Thursday, 14 March 2002, Emiliano wrote: Richard Stallman wrote: I think these issues should be judged by the substance of the requirement rather than by the legal hook which is used to impose it. For instance, a requirement to make source available to users is substantively a requirement of distribution rather than a restriction on use. At present we are planning to try to handle the ASP problem in the GPL through a limitation on a certain kind of modification--that you can't delete or disable a command that lets the user download source (if the program has one to start with). Lawyers we have consulted think that will work. Any indication on when this would be available? Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3 -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
On Friday 15 March 2002 01:06 am, Emiliano wrote: Whether or not you feel that this is or isn't a good way to deploy applications is entirely irrelevant to me. I have an application to release; if I can't release it under an open source license because these types of apps are deemed irrelevant by the OSI then so be it. Closed source it is. I still have hopes for the amended GPL or a variant of the APSL or W3C license, though. I was not judging any type of application. Instead I was prognosticating from my own particular perspective. No offense was intended and I hope you don't take it as such. I do foresee a need for web services and other forms of centralized processing. But I still doubt that they will become the dominant mode of application computing. I get the impression that there are one or two people out there (not you) who are in a state of urgency over this issue because they see an immanent return of the dumb/thin client. I don't like decisions to be made in panic mode, so I was trying to explain in my own not-to-brilliant way why I didn't think it's going to happen. -- David Johnson ___ http://www.usermode.org pgp public key on website -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Richard Stallman writes: The reason we've decided that this ASP requirement is legitimate is that it is a matter of requiring making the modified source code available in a case of public use. It extends existing GPL requirements coherently to a new scenario of usage. We've never intentionally approved use-restricted licenses before. Several non-free licenses have been put before us, which had use restrictions (using the same justification you are currently using). We didn't approve those licenses. I will vote against approving the GPLv3 if it imposes restrictions on users. It would be wrong to require publication of modified versions that are used privately, but inviting the public to use a server is not private use. I'm not sure that the GPL-using community is going to agree with you on this. For example, if someone decides to distribute Linux under the GPLv3, and someone else runs it on a server, are they bound by the GPLv3 to become a Linux distributor? I haven't seen it, so I'm just speculating. -- -russ nelson http://russnelson.com | Crypto without a threat Crynwr sells support for free software | PGPok | model is like cookies 521 Pleasant Valley Rd. | +1 315 268 1925 voice | without milk. Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Richard Stallman writes: A simple example: it is totally trivial on Windows to build a 'service' from a DLL, exposing its entire interface. This would be running as a separate executable, but would look like a regular library to any windows program. The FSF's position is that the GPL applies to any programs which are designed to link with that DLL, that this is legally equivalent to statically linking them. That is your position because it must be your position, not because there is any legal viability to the argument. How do you know that such an intention was there? Perhaps the program was designed to link with another, compatible non-GPL DLL, and mere compatibility allows it to link with the GPL DLL? I really don't know what a judge would say, and I expect that you don't know either. -- -russ nelson http://russnelson.com | Crypto without a threat Crynwr sells support for free software | PGPok | model is like cookies 521 Pleasant Valley Rd. | +1 315 268 1925 voice | without milk. Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Russell Nelson [EMAIL PROTECTED] writes: Richard Stallman writes: It would be wrong to require publication of modified versions that are used privately, but inviting the public to use a server is not private use. I'm not sure that the GPL-using community is going to agree with you on this. For example, if someone decides to distribute Linux under the GPLv3, and someone else runs it on a server, are they bound by the GPLv3 to become a Linux distributor? I haven't seen it, so I'm just speculating. As I understand it, they would only be required to not remove a ``fetch source here'' button. So the issue then becomes whether the hypothetical GPLv3 Linux has a ``fetch source here'' button; I've never seen a version of Linux with such a button. Ian -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Emiliano scripsit: Well, I could answer that in two, conflicting ways. If distribution becomes irrelevant, the spirit of the GPL in that respect is obsolete, isn't it? I don't see how that could happen, unless bandwidth (including the last mile) becomes too cheap to meter. I don't think that's relevant. Web Services will, by definition, be served this way. That they may not be available (or viable) to some is a different issue. My point was that distribution will not become irrelevant just because of the Web Services hype. In the reasonably foreseeable future, most programs will continue to execute locally because of bandwidth constraints. -- John Cowan [EMAIL PROTECTED] http://www.reutershealth.com I amar prestar aen, han mathon ne nen,http://www.ccil.org/~cowan han mathon ne chae, a han noston ne 'wilith. --Galadriel, _LOTR:FOTR_ -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
John Cowan wrote: It's not as bad as that: copyright law lets you do the things you want. You may or may not pass the OSD, though. I'd prefer to have it pass the OSD. I prefer more, though, to release the sources. Releasing the sources is bound by some restrictions that may or may not become OSD-conform; it's not at all likely that they'll be anything like the GPL for reasons outlined below. Here's a sketch of a license (IANAL, TINLA): 1) You, the recipient of this software, may make verbatim copies of the software and distribute them to anyone. 2) You may make derivative works from the software, provided that you make the source code of your derivative work available either by allowing public downloading or by sending it to anyone on request. You may distribute such derivative works in any way you see fit. Your derivative works must be released under this license. But what if I obtain a copy, make private mods, and offer them for download under the URL http://www.mysite.com/discontinued/oos+mods.html This URL, furthermore, is only reacheable by an arcane path through my website (training/contacts/personal pages/generic/products/legal/old site/images/trial), 10 layers deep. I also do not post notification of this URL in we forms displayed by the web app. Is it publicly available? Yes. Will anyone know? No, since I'm not required to advertize the existense of the download and or its license. People specifically asking for it will get the URL, but how will people know to ask? The hook here is that private modifications are flatly not allowed; the minute you make a modification, you commit to making your modified source available to all. That would be good for us. We'd also like there to be an obvious way for the service-user to find a copy of our own license. Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
I think these issues should be judged by the substance of the requirement rather than by the legal hook which is used to impose it. For instance, a requirement to make source available to users is substantively a requirement of distribution rather than a restriction on use. At present we are planning to try to handle the ASP problem in the GPL through a limitation on a certain kind of modification--that you can't delete or disable a command that lets the user download source (if the program has one to start with). Lawyers we have consulted think that will work. -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Richard Stallman wrote: I think these issues should be judged by the substance of the requirement rather than by the legal hook which is used to impose it. For instance, a requirement to make source available to users is substantively a requirement of distribution rather than a restriction on use. At present we are planning to try to handle the ASP problem in the GPL through a limitation on a certain kind of modification--that you can't delete or disable a command that lets the user download source (if the program has one to start with). Lawyers we have consulted think that will work. Any indication on when this would be available? Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Emiliano writes: I see the issues this brings, but my immediate choices are 1) keep the source closed, or 2) release the sources under conditions that give the users of the service that my software will provide the same freedoms as the entity running that software for them has. You can certainly do this, but you'll need some component which is not free or open source software. Something like this: #!/bin/sh # Copyright 2002, Emiliano. All rights reserved. cat EOF This is a license. If you unpack the enclosed tarball, you have agreed to the license. [ license terms ] EOF echo -n Do you agree to the license (y/n)? read answer case $answer in Y|y|yes) echo -n unpacking the tarball tr a-mn-z n-za-m tarball.tar.gz | tar xpf - echo ;; *) echo Okay, then you have no permission to unpack the enclosed tarball. esac -- -russ nelson http://russnelson.com | Crypto without a threat Crynwr sells support for free software | PGPok | model is like cookies 521 Pleasant Valley Rd. | +1 315 268 1925 voice | without milk. Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
John Cowan writes: It's not as bad as that: copyright law lets you do the things you want. No it doesn't: 2) You may make derivative works from the software, Copyright law already gives them permission to make derivative works. They just can't *copy* them. But they don't want to, and aren't going to. -- -russ nelson http://russnelson.com | Crypto without a threat Crynwr sells support for free software | PGPok | model is like cookies 521 Pleasant Valley Rd. | +1 315 268 1925 voice | without milk. Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Richard Stallman writes: I think these issues should be judged by the substance of the requirement rather than by the legal hook which is used to impose it. For instance, a requirement to make source available to users is substantively a requirement of distribution rather than a restriction on use. At present we are planning to try to handle the ASP problem in the GPL through a limitation on a certain kind of modification--that you can't delete or disable a command that lets the user download source (if the program has one to start with). Lawyers we have consulted think that will work. I doubt we would approve such a license. We refused to approve Larry McVoy's Bitkeeper License precisely because it had a limitation on a certain kind of modification. The public version of Bitkeeper had a license term that didn't let you remove or modify a module which forced you to publish all modifications on a public bitkeeper server. If you didn't want to do that, you would have to license the code. -- -russ nelson http://russnelson.com | Crypto without a threat Crynwr sells support for free software | PGPok | model is like cookies 521 Pleasant Valley Rd. | +1 315 268 1925 voice | without milk. Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
We are helping a third party to incorporate our proposed solution to the ASP problem in its own modified GPL for release very shortly (probably within days). Legal work is complete and only some public information documents are not yet final. Use in an FSF-approved third-party license will be followed by inclusion of the term in a draft of GPLv3 that we hope to release for public comment later this year. Our change is not a state secret. Richard has described how it works. If you want to make a new web application's code fully free in ASP use, you release your first version with a download server source button prominently located on pages every user sees. The new license provision extends GPLv2's rule that you cannot remove the copyright notice display from an interactive program to say that if you modify the code you cannot remove the download server source button and functionality. In this way, license terms constrain only modification, in a fashion completely compliant with FSD and OSD. Existing applications are unaffected. New applications and new versions of existing applications can be written so that someone who offers application services must also distribute the conforming source code to all users. Because we have a partner here, I don't want to prerelease the text of its license. If you plan a release within a week, I will be happy to give you my earlier drafts of the language, and help you to make a modified GPL of your own. The whole matter will be completely public so soon, however, that you will almost certainly prefer to wait for announcement. On Thursday, 14 March 2002, Emiliano wrote: Richard Stallman wrote: I think these issues should be judged by the substance of the requirement rather than by the legal hook which is used to impose it. For instance, a requirement to make source available to users is substantively a requirement of distribution rather than a restriction on use. At present we are planning to try to handle the ASP problem in the GPL through a limitation on a certain kind of modification--that you can't delete or disable a command that lets the user download source (if the program has one to start with). Lawyers we have consulted think that will work. Any indication on when this would be available? Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Russell Nelson wrote: You can certainly do this, but you'll need some component which is not free or open source software. Something like this: #!/bin/sh # Copyright 2002, Emiliano. All rights reserved. cat EOF This is a license. If you unpack the enclosed tarball, you have agreed to the license. [ license terms ] EOF echo -n Do you agree to the license (y/n)? read answer case $answer in Y|y|yes) echo -n unpacking the tarball tr a-mn-z n-za-m tarball.tar.gz | tar xpf - echo ;; *) echo Okay, then you have no permission to unpack the enclosed tarball. esac A: accepts license, tells B: 1. download tarball 2. do tr a-mn-z n-za-m tarball.tar.gz | tar xpf - Has A violated the license? That would seem weird, because it pertains clear written instructions (even if those instructions are in fact a shell script). And since B didn't agree to anything, he isn't bound by the 'click-wrap' license. Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Russell Nelson scripsit: Copyright law already gives them permission to make derivative works. That is not so. One may not make a derivative work without the permission of the copyright holder. For example, one may not translate a book into a different language without such permission, even an unpublished translation. If you distribute the translation, the breach is compounded, but even if you don't, it is a breach. They just can't *copy* them. But they don't want to, and aren't going to. The term copyright is not restricted to the exclusive right to make copies (that is, to prevent others from making them). It includes also the exclusive rights to distribute, to publicly display, to publicly perform, *and* to make derivative works based on the copyrighted work. -- John Cowan [EMAIL PROTECTED] http://www.reutershealth.com I amar prestar aen, han mathon ne nen,http://www.ccil.org/~cowan han mathon ne chae, a han noston ne 'wilith. --Galadriel, _LOTR:FOTR_ -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Emiliano writes: A: accepts license, tells B: 1. download tarball 2. do tr a-mn-z n-za-m tarball.tar.gz | tar xpf - Has A violated the license? That would seem weird, because it pertains clear written instructions (even if those instructions are in fact a shell script). And since B didn't agree to anything, he isn't bound by the 'click-wrap' license. The dragon attacks: Ask your lawyer to include terms in the license that prohibit disclosure of the unpacking method. The license will also have to prohibit distribution of the enclosed tarball. -- -russ nelson http://russnelson.com | Crypto without a threat Crynwr sells support for free software | PGPok | model is like cookies 521 Pleasant Valley Rd. | +1 315 268 1925 voice | without milk. Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
RE: OSD modification regarding what license can require of user
RN Copyright law already gives them permission to make derivative works. JN That is not so. One may not make a derivative work without the permission of the copyright holder. For example, one may not translate a book into a different language without such permission, even an unpublished translation. If you distribute the translation, the breach is compounded, but even if you don't, it is a breach. /John Cowan This is a good point, but it need to be developed. Let me add that once permission is given, Copyright Law can give to second author of derivative work the SAME rights than the first author, on their respective versions. How can we understand these contradictions ? I want to bring to your attention these derivative rights because these 2 concepts we need to distinguish: derivative rights, and distribution rights (right to copy). These rights are not exactly the same, and the implications are different. The assertion that respective rights on derivative are EQUALS to those of the author can be found in Treaties (Berna and WIPO developments) http://clea.wipo.int/lpbin/lpext.dll/clea/LipEN/46e4b/46e4c?f=file%5Bdocumen t.htm%5D#JD_388eb In particular: Article 2, alinea (3). Most countries adhere to it (including US, of course). [You can notice this article apply to artistic works. I do not want to discuss if software is an artistic work. I only notice that software licenses invoke IP Laws to implement their open-protection scheme.] If we say in short that first and second author have the same rights, somebody can think this is anarchy. This is not so. Copyright Law (WIPO Treaties) introduces a powerful protection scheme which can be rephrased as: Rights of licensees on their respective work will be the SAME, _without prejudice_ of the rights of the author, AND ONLY IF author authorizes the derivation. (see for example: art.8 for translations, and art 12 for modifications) IP Laws define this protection scheme as moral rights. In US, the usual word used is copyright. But this is the same concept and all these expressions (in each legal tradition) are used in conformity with this WIPO Convention. You can use the word tutelage, too, as in French tradition. There is no country differences under IP Laws. This universal, underlying IP mechanism allows for example all licenses (OS or non OS) to work in a strong manner. While the author does not allow the derivation, second authors have no rights. Period. But as soon as the author allows derivation, second authors are granted the same rights than author on his/her respective version. Again, it is very important to understand that the rights given to second authors on their derivative versions are granted by Laws, not by authors. Of course, Country Laws can varie on details, they are rephrased in different languages and follow different traditions, but all are compliant with the principle of moral rights/copyright. This is the Berna Convention. The true difference between OS and non-OS licenses is that OS licenses do not require user identification nor payment from licensees, while non-OS licenses requires payment, and if they can (as we see in most Web sites, today) they require also the identification of licensees... :-D) If asked, I should insert this rule as #0 in OSD : do not require identification of licensee. Based on the above, I want to underline some of the reasons that cause that OS mechanism fails under some circumstances. Actually the OS mechanism is based on the idea that using the IP protection given to authors on distribution, we can also protect derivatives versions of work. This solution fails (or will fail) because the rights on derivative works are not under control of first author. Everyone can derive a work, as long as it do not try to distribute his version. Authors cannot grant rights for derivation because this right is already given by laws. There is no society nor progress without the right to derive from Sources, books, ideas, or goods. Rights for distribution and for derivation do not come from the same source. Rights to distribute a work belong to copyholder: they can be limited on time, space and quantity. Rights on derivation work cannot be limited: they belong entirely to second authors. A copyholder cannot define a license that make prejudice to the rights owned by second authors on their respective versions. In fact, a version-holder can legally forbid the distribution of its own derivative work, based on the _same rights_ than allows the original-holder to forbid the derivation. A license which grant derivative rights is almost a non-sense. Authors can *forbid* the distribution of the derivative work, but they cannot forbid the freedom of people to create derivations. Authors cannot claim for rights on derivative works, because these rights belongs to the author of the version. This is the meaning of some of the court examples that some of you are describing. For all these
Re: OSD modification regarding what license can require of user
On Thursday 14 March 2002 03:20 am, John Cowan wrote: I don't think that's relevant. Web Services will, by definition, be served this way. That they may not be available (or viable) to some is a different issue. My point was that distribution will not become irrelevant just because of the Web Services hype. In the reasonably foreseeable future, most programs will continue to execute locally because of bandwidth constraints. It's more than bandwidth contraints. A future where all applications run on remote servers and all the clients are thin is foreseeable, but not likely. We had that once, back in the days when RMS decided not to place restrictions on usage but only on distribution. We've moved beyond that. We now realize that it is more efficient to have a thousand CPUs process a thousand applications than to have a single centralized CPU process those same thousand applications. Personally I don't think this latest big push towards thin clients is going to be any more successful than the earlier pushes. Some of it will of course happen, but I don't expect it to be the common mode of executing applications. I do foresee distributed processing. But this involves distribution, so no problem. But centralized processing has come and gone. -- David Johnson ___ http://www.usermode.org pgp public key on website -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
I.R.Maturana scripsit: This is a good point, but it need to be developed. Let me add that once permission is given, Copyright Law can give to second author of derivative work the SAME rights than the first author, on their respective versions. Correct. This is a point that is often not understood. Rights of licensees on their respective work will be the SAME, _without prejudice_ of the rights of the author, AND ONLY IF author authorizes the derivation. Exactly so. IP Laws define this protection scheme as moral rights. In US, the usual word used is copyright. But this is the same concept and all these expressions (in each legal tradition) are used in conformity with this WIPO Convention. Moral rights are a bit different (things like the right to be known as the creator, the right to prevent mutilation of the work, etc.), and the U.S. applies them only to visual works. There is no country differences under IP Laws. There are differences, but the ones you mentioned are not among them, true. What counts as fair use, for example, varies from country to country. This solution fails (or will fail) because the rights on derivative works are not under control of first author. Everyone can derive a work, as long as it do not try to distribute his version. No, this contradicts what you say above and is not true under either U.S. law or the Berne convention. The *exclusive* right to create derivative works is one of the copyright owner's rights. For example, you cannot translate a novel from French to English without the consent of the copyright owner (initially the author) of the French version. Nor can you make a book into a film without the agreement of the copyright author. Authors cannot grant rights for derivation because this right is already given by laws. There is no society nor progress without the right to derive from Sources, books, ideas, or goods. That is not correct. Rights for distribution and for derivation do not come from the same source. They do: the Berne Convention and the various national copyright acts. Rights on derivation work cannot be limited: they belong entirely to second authors. A copyholder cannot define a license that make prejudice to the rights owned by second authors on their respective versions. In fact, a version-holder can legally forbid the distribution of its own derivative work, based on the _same rights_ than allows the original-holder to forbid the derivation. This part is correct: once the derivative work is created *under license*, the deriving author can exclude the original author. But that original license = permission is required, and if it is denied, the derivative author has no right to create the derivative work. A license which grant derivative rights is almost a non-sense. Authors can *forbid* the distribution of the derivative work, but they cannot forbid the freedom of people to create derivations. That is not the case. Authors cannot claim for rights on derivative works, because these rights belongs to the author of the version. That is true. -- John Cowan [EMAIL PROTECTED] http://www.reutershealth.com I amar prestar aen, han mathon ne nen,http://www.ccil.org/~cowan han mathon ne chae, a han noston ne 'wilith. --Galadriel, _LOTR:FOTR_ -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
John Cowan writes: Russell Nelson scripsit: Copyright law already gives them permission to make derivative works. That is not so. http://cr.yp.to/softwarelaw.html They just can't *copy* them. But they don't want to, and aren't going to. The term copyright is not restricted to the exclusive right to make copies (that is, to prevent others from making them). It includes also the exclusive rights to distribute, to publicly display, to publicly perform, *and* to make derivative works based on the copyrighted work. An ASP desires to do none of these. -- -russ nelson http://russnelson.com | Crypto without a threat Crynwr sells support for free software | PGPok | model is like cookies 521 Pleasant Valley Rd. | +1 315 268 1925 voice | without milk. Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Russell Nelson scripsit: http://cr.yp.to/softwarelaw.html I don't find DJB's claim that private modifications are allowed credible. His quotation from the CONTU Final Report is misleadingly selective. A fuller quotation states: Thus a right to make those changes necessary to enable the use for which it was both sold and purchased should be provided. The conversion of a program from one higher-level language to another to facilitate use would fall within this right, as would the right to add features to the program that were not present at the time of rightful acquisition. Note the conditional: should be provided. There is no evidence that the right (implicitly a new one) was in fact added to 17 U.S.C. perform, *and* to make derivative works based on the copyrighted work. An ASP desires to do none of these. The concern here is with an ASP making private modifications to its version of the program. Those private modifications create a derivative work. -- John Cowan [EMAIL PROTECTED] http://www.reutershealth.com I amar prestar aen, han mathon ne nen,http://www.ccil.org/~cowan han mathon ne chae, a han noston ne 'wilith. --Galadriel, _LOTR:FOTR_ -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Bruce Perens scripsit: The assignment is to propose a modification to the OSD to make explicit the implicit prohibition in the OSD on odious requirements on the user. The examples of odious user requirements that are brought up most often are badgeware (if you use this software, put my icon on your home page) and spyware (the Bitkeeper reporting feature), but there have been others. Mooseware, e.g. But I don't see why the Apache license isn't badgeware. Nintendo claimed that the Goloob cartrige was a derivative work, Goloob disagreed. The judge found for Goloob. The judge made a point of noting that this case should not be considered to be definitive. This sounds on all fours with the claim that patches are not derivative works (with which I agree): they make only fair use of the original. Let's consider also the ASP problem. Somebody makes extensive changes to a GPL work, and deploys that work as a service, perhaps via .NET, rather than distributing the work. This circumvents the GPL because the GPL terms activate upon distribution. I don't understand how this breaches the spirit of the GPL any more than providing ASP-style access to the unmodified work does (i.e. not at all). If you are free to make private mods to GPLed programs for your own use, why not for others' use? This is just timesharing under a new name. -- John Cowan [EMAIL PROTECTED] http://www.reutershealth.com I amar prestar aen, han mathon ne nen,http://www.ccil.org/~cowan han mathon ne chae, a han noston ne 'wilith. --Galadriel, _LOTR:FOTR_ -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
John Cowan wrote: Let's consider also the ASP problem. Somebody makes extensive changes to a GPL work, and deploys that work as a service, perhaps via .NET, rather than distributing the work. This circumvents the GPL because the GPL terms activate upon distribution. I don't understand how this breaches the spirit of the GPL any more than providing ASP-style access to the unmodified work does (i.e. not at all). Because the unmodified form is available to the service users, and they could escape if they wanted to and set up a private implementation. Not so with provider-private mods. If you are free to make private mods to GPLed programs for your own use, why not for others' use? This is just timesharing under a new name. I, for example, write an online questionaire application that I'm considering releasing under on OSD compliant license. The main thing holding me back so far is that I'm still looking for a license that allows users their freedom, but prevent direct competitors to my bussiness taking what's mine, _making private mods_, and competing with my own service (which is just fine with me up to here) without granting their users the freedom I've given them, or even an acknowledgement of my work or a notice of where to get a copy of their own, because the service-users will never get to see the source and/or the license. And I'm not fine with that. Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
John Cowan: I don't understand how this breaches the spirit of the GPL any more than providing ASP-style access to the unmodified work does (i.e. not at all). If you are free to make private mods to GPLed programs for your own use, why not for others' use? This is just timesharing under a new name. Well, I could answer that in two, conflicting ways. If distribution becomes irrelevant, the spirit of the GPL in that respect is obsolete, isn't it? On the other hand, Richard treats this as a privacy issue. I contend that a publicly performed GPL work with a private implementation does indeed contradict the spirit of the GPL. Then, you get into the question of what is the entity in which privacy applies? In the GPL, it is the entity within which there is no need to distribute. This has always been sort of vague to me, because it's not clear if a corporation and all of its employees are a single entity, or if distribution takes place between employees, or between employees and the corporation. Then, bring in complications like consultants and companies working under contract but not part of the same legal entity as the corporation. Bruce -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Bruce Perens scripsit: Well, I could answer that in two, conflicting ways. If distribution becomes irrelevant, the spirit of the GPL in that respect is obsolete, isn't it? I don't see how that could happen, unless bandwidth (including the last mile) becomes too cheap to meter. On the other hand, Richard treats this as a privacy issue. I contend that a publicly performed GPL work with a private implementation does indeed contradict the spirit of the GPL. It's not at all clear to me that when I send you bits, you massage them on your own computer, and you send me different bits back, that this constitutes a public performance of anything. It sounds a lot more like research or consulting. Suppose A publishes a GPLed book describing some arcane subject, and B obtains a copy of it. C now mails questions to B along with payment, and B answers the questions out of the book and mails back the replies. In principle, C could read the book himself, but may not have the time or desire.) Surely A's rights are not impinged on here? Are things different if B adds his own marginal notes to the book? Is B really required by (the spirit of) the GPL to make those notes available to C? This has always been sort of vague to me, because it's not clear if a corporation and all of its employees are a single entity, or if [...] I agree that all this is sticky. -- John Cowan [EMAIL PROTECTED] http://www.reutershealth.com I amar prestar aen, han mathon ne nen,http://www.ccil.org/~cowan han mathon ne chae, a han noston ne 'wilith. --Galadriel, _LOTR:FOTR_ -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
From: John Cowan [EMAIL PROTECTED] I don't see how that could happen, unless bandwidth (including the last mile) becomes too cheap to meter. Think about think clients and internet appliances. If a lot of people go to thin clients because they want to oursource their system administration, then the paradigm changes. It's not at all clear to me that when I send you bits, you massage them on your own computer, and you send me different bits back, that this constitutes a public performance of anything. You would not doubt that it was a public performance if those bits were a television broadcast. Suppose A publishes a GPLed book describing some arcane subject, and B obtains a copy of it. C now mails questions to B along with payment, and B answers the questions out of the book and mails back the replies. In principle, C could read the book himself, but may not have the time or desire.) Surely A's rights are not impinged on here? If B cut and pasted the answers _directly_ from the book, and did so to an extent greater than the simple occassional quoting within a larger work allowed as fair use, B would indeed be conveying a copyrighted work to C, and the license would apply. If B, on the other hand, conveyed the _knowledge_ rather than its representation as created by A, there would not be any conveyance of A's copyrighted work. But that's not really what we're talking about here. B is not answering C based on mere knowledge of the output of A's program. B is providing C with a means of accessing A's program as if C was the party to whom the work had been licensed. Although C can't see the source or binary code of the copyrighted work, it is executed upon his behalf, at his command, and he gains the benefit of its execution. Are things different if B adds his own marginal notes to the book? There is a boundary to fair use, I don't think marginalia would fit one within it. Is B really required by (the spirit of) the GPL to make those notes available to C? I submit that this is the case. But the implementation of copyright law and the current text of the GPL both leave a lot to be desired when this sort of question comes up. Thanks Bruce -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
On Wed, Mar 13, 2002 at 10:24:02PM +, Thorsten Glaser wrote: Or A, to look from a different side on this? No. The terms of the GPL don't require you to give source code to the person to whom you distribute the binary. If you fail to do so, the copyright holder can sue you for infringement, but unless you manage to give the copyright holder a binary, they can't compel you to give the source code to them. Bruce -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Bruce Perens scripsit: It's not at all clear to me that when I send you bits, you massage them on your own computer, and you send me different bits back, that this constitutes a public performance of anything. You would not doubt that it was a public performance if those bits were a television broadcast. No, but that has to do with the copyright status *of the bits sent back*, not at all with the copyright status of the program doing the massaging (unless it is Bison-type and includes part of itself in the output). If B cut and pasted the answers _directly_ from the book, and did so to an extent greater than the simple occassional quoting within a larger work allowed as fair use, B would indeed be conveying a copyrighted work to C, and the license would apply. But that is the Bison case. (Yes, I know that the Bison skeleton license has been changed.) To be concrete, suppose I provide a fast grepping service. You send me a regex and some URLs, and I use GNU grep to to send you specific parts of the documents specified by the URLs. We will neglect whether the copyright of the document's author is breached by this. If I make private modifications to GNU grep to improve it, I don't see that the GPL (in letter or spirit) requires me to redistribute those modifications. I am *using* grep to provide a service. By the same token, when I send a form to a Web site, the GPLed program Perl may be executed on my behalf. If Perl has some local hacks, am I entitled to have those hacks (the copyright in which is owned by the site's programmer) to be sent to me? That sounds like forced distribution to me. (Again, let us neglect Perl's alternative licensing.) -- John Cowan [EMAIL PROTECTED] http://www.reutershealth.com I amar prestar aen, han mathon ne nen,http://www.ccil.org/~cowan han mathon ne chae, a han noston ne 'wilith. --Galadriel, _LOTR:FOTR_ -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
CORRECTION: Re: OSD modification regarding what license can require of user
Darn. I garbled. Delete the don't. The terms of the GPL require you to give source code to the person to whom you distribute the binary. And nobody else. The copyright holder can sue for infringement but can't compel the infringer to give the copyright holder a copy of the source code. He can compel the infringer to give the source code to the person to whom he gave the binary. Hopefully this parses better. Thanks Bruce -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
From: John Cowan [EMAIL PROTECTED] To be concrete, suppose I provide a fast grepping service. Grep is an over-simple case, which might lead you to trivialize the problem. Consider Evolution, OpenOffice, or GNU Emacs. Postulate that someone makes a way for somebody to use one of those programs as if it were running natively on their computer, without ever activating the distribution terms of the GPL. And that same someone makes significant enhancements which he does not disclose in either source or binary form. Consider this from the perspective of the creator of Evolution, OpenOffice, or GNU Emacs. This person has put an immense amount of work out in the public with the expectation that improvements that are widely used would be distributed, and thus would be returned to them. And they aren't. Is this fair to them? I contend that this sort of activity should be placed outside of the covenant represented by the GPL. Richard and Eben don't necessarily agree with me - yet. One thing that we learned in deploying the OSD is that the world would surprise us, and that the language would end up being more important than we thought. I don't want to be surprised as much the next time around. Thanks Bruce -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
On Wed, 13 Mar 2002, John Cowan wrote: To be concrete, suppose I provide a fast grepping service. You send me a regex and some URLs, and I use GNU grep to to send you specific parts of the documents specified by the URLs. We will neglect whether the copyright of the document's author is breached by this. If I make private modifications to GNU grep to improve it, I don't see that the GPL (in letter or spirit) requires me to redistribute those modifications. I am *using* grep to provide a service. Or take it even further, and consider the case where people use privately modified software to provide some other service. For example, people send me images via email or Zip disk, and I use a modified Gump to create some really nice effects and email/Zip disk the modified images back. Now it's not even a web service, in that it's not automatically called, but it's still used as a crucial step in the delivery of a service. Would I be (legally, morally, ethically) required to share the code to my modifications to Gump? Is the fact that it's automated in John's example and not automated in mine relevant? (For the purposes of this example I'm assuming Gump has a GPL license and doesn't make exceptions for plug-ins) Brian -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
On Wed, 13 Mar 2002, Bruce Perens wrote: Consider Evolution, OpenOffice, or GNU Emacs. Postulate that someone makes a way for somebody to use one of those programs as if it were running natively on their computer, without ever activating the distribution terms of the GPL. And that same someone makes significant enhancements which he does not disclose in either source or binary form. Yep, like making it available through VNC, for example. A very clear violation of the spirit of the GPL; but the grey area between this and the examples in the earlier messages seems very hard to divide between clear and non. Brian -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
From: Brian Behlendorf [EMAIL PROTECTED] Yep, like making it available through VNC, for example. A very clear violation of the spirit of the GPL; I'm glad you agree. but the grey area between this and the examples in the earlier messages seems very hard to divide between clear and non. It seems that the role of the user is key. Is the user running the program? Thanks Bruce -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Emiliano writes: their users the freedom I've given them, or even an acknowledgement of my work or a notice of where to get a copy of their own, because the service-users will never get to see the source and/or the license. And I'm not fine with that. That Way Lies Dragons. Since the service provider is not copying the modified work when they let someone use it, the copyright holder (henceforth, 'you') needs more than copyright permissions to solve this problem. You need to get people to agree to distribute modifications[1], and you need to get them to agree to penalties if they refuse. Such an agreement would have to be enforcable in a court of law. As such, it would be a contract. A contract has to be negotiated. It cannot be applied unilaterally. The negotiation might be as simple as an offer and an acceptance. How do you distribute Open Source software in such a manner that the user can only access it after seeing an offer and performing an act which can only be interpreted as acceptance of that offer? [1] I don't see how such a license could be a free software license either, given what RMS has said about privacy in: http://www.gnu.org/philosophy/free-sw.html -- -russ nelson http://russnelson.com | Crypto without a threat Crynwr sells support for free software | PGPok | model is like cookies 521 Pleasant Valley Rd. | +1 315 268 1925 voice | without milk. Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
RE: OSD modification regarding what license can require of user
Hello Sorry for my English, but let me try to offer another perspective. Consider this from the perspective of the creator of Evolution, OpenOffice, or GNU Emacs. This person has put an immense amount of work out in the public with the expectation that improvements that are widely used would be distributed, and thus would be returned to them. And they aren't. Is this fair to them? I contend that this sort of activity should be placed outside of the covenant represented by the GPL. Richard and Eben don't necessarily agree with me - yet. Let us make a difference between the spirit of the work and the use of the work. Then let me try to extract some solution. The definition of the purpose of the work belongs to the author. On first side, moral authority (or tutelage) of the author is the source of the actual rights granted by any license. On the other side, the possible use of the work is always defined by a license. When the author is and individual, the right to define the license belongs to the individual. When the author is member of a corporation, (if the work has been developed, or paid by a corporation), the right to define a license belongs to the corporation. In both cases, Copyright/Intellectual Property (IP) Laws only define who is the copy-holder, and which are the rights that the copyholder can grant. The extended use of expressions like Open license or restricted license is somehow confusing. Open/Restricted refers to the purpose of author, not to the license itself. The only true open license is a public domain license. Strictly speaking, all OS licenses are always restrictives. The point here, is to understand that the mechanism in OS licenses is based on a legal formalism. Theses licenses are legally strong because they are based on the same spirit of private interest than non-OS licenses. This formalism works well because copyright and IP Laws are based on the supposition that authors are always restrictive. Laws assume that if an author creates some work, he will never want to publish nor grants user rights without garantees. Therefore Laws are designed to offer protection to the restrictions that authors want to apply. The good idea of GPL was to use this legal protection of Laws to enforce the free use, derivation and distribution of work, instead of a restrictive use. The issue to this approach is that Laws are still based on a restriction logic. This way, any open licence can expect to reproduce indefinitely some contradiction like the Achille Paradox. The basic contradiction is that you cannot define the public interest of OS authors as a result of the private interest definition of the IP Laws. The private interest can only be defined inside the boundaries of the public interest. (For example, antitrust laws apply against MS because private interest cannot destroy the market -- fair market is of public interest.) Now, let me try to define the opposite mechanism. We start on the idea that authors have a public interest and we go ahead, until we find the connexion with private IP laws. The mechanism is as follow: 1) Any agreement between two individuals defines some legal space Define this agreement as follow: The authors grants rights to anybody to access, derive, modify copy and translate his work, ONLY ON THE SPACE DEFINED BY THIS AGREEMENT The condition is that licensee MUST allow the access, derivation, modification and translation of his derivative work ON THE SAME SPACE DEFINED BY THIS AGREEMENT, with no fee nor identification of licensee. Name this resulting space : the Internet 2) Suspend any right of use of the work outside this space. That is, copy, modify, translation etc.. is forbidden if people cannot access the copy or derivative inside Internet, Strictly speaking, there is no right to suspend, because authors have granted no rights at all. Because no rights have been granted outside internet, all the rights remains to author. Therefore, outside this space, outside Internet, authors holds all the rights. He can earn money, publish, distribute his work under the Laws of each country. He can even prohibit the use to copy, distribute or use the work outside Internet. 3) Assert that the copyrights of the licensees will be equals to the rights of the author, on their respectives versions. Note this last assertion is given by IP Laws (not by license) This assertion exists in WIPO treaties and all country Laws. This way, this licence becomes a legal document valid on all countries, under all Laws, and if you translate it, this license become valid in all languages. You have now connected the public will of open minded authors and licensees with country Laws, and you can claim for their protection against closed minded authors, in each country, in each language, under all courts. Conclusion. Given the handshake of open-minded developers, let us define a legal
Re: OSD modification regarding what license can require of user
Bruce wrote (in part) So, what if it turns out that the present GPL doens't hold up with regard to dynamic linking? Some future version of the GPL might have to place a constraint on the user regarding combination of works on the user's system that would, if it were distributed in that form, be considered a derived work. I think that should be allowed by the OSD. Did you somehow mean constraint on the publisher regarding combination of works on the user's system? Thus, there seem to be a few sorts of requirement on the user that I think _should_ be permitted by the OSD. All of them are intended to further the goals of Open Source or Free Software. freedom and requirement are direct opposites. Imposing requirements may further the goals of Open Source, (whatever they are, I thought they were freedom.) Imposing requirements on users is unlikely to further freedom, by definition. -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Bruce wrote (in part) So, what if it turns out that the present GPL doens't hold up with regard to dynamic linking? Some future version of the GPL might have to place a constraint on the user regarding combination of works on the user's system that would, if it were distributed in that form, be considered a derived work. I think that should be allowed by the OSD. On Thu, Mar 14, 2002 at 01:05:34AM -0500, Forrest J. Cavalier III wrote: Did you somehow mean constraint on the publisher regarding combination of works on the user's system? If we really _can_ implement this as a constraint on the distributor, rather than the user, that would leave us in _much_ better shape. Would you like to float some sample legal language to implement that? I have access to a legal staff to review it. If we can assure ourselves that it's possible, that leaves indemnification and the ASP problem. Indemnification is a must, IMO. The ASP problem, in contrast, is a wish-list item we could lose. freedom and requirement are direct opposites. Of course. I hope you accept that we _don't_ live in the libertarian utopia. Thus, some turning of the dominant legal structure upon its head is necessary. Our only avenue for doing that is the creative use of legal restriction. Thanks Bruce -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Bruce Perens wrote: John Cowan: I don't understand how this breaches the spirit of the GPL any more than providing ASP-style access to the unmodified work does (i.e. not at all). If you are free to make private mods to GPLed programs for your own use, why not for others' use? This is just timesharing under a new name. Well, I could answer that in two, conflicting ways. If distribution becomes irrelevant, the spirit of the GPL in that respect is obsolete, isn't it? On the other hand, Richard treats this as a privacy issue. I contend that a publicly performed GPL work with a private implementation does indeed contradict the spirit of the GPL. Then, you get into the question of what is the entity in which privacy applies? In the GPL, it is the entity within which there is no need to distribute. This has always been sort of vague to me, because it's not clear if a corporation and all of its employees are a single entity, or if distribution takes place between employees, or between employees and the corporation. Then, bring in complications like consultants and companies working under contract but not part of the same legal entity as the corporation. Bruce As to the question of the scope of a corporate entity, the MPL uses a control standard often found in corporate documents, securities documents and statutes. The language is the sort everyone hates -- clearly written by lawyers for laawyers. I've included it below. But it is in general use, is generally understood by lawyers and judges who may someday look at this, and so won't need to develop its own case law and understanding. There is a couple of other standards in use for determinging control but the one in the MPL seemed the best, at least when I researched this last. Mitchell *++ 1.12. You'' (or Your) * means an individual or a legal entity exercising rights under, and complying with all of the terms of, this License or a future version of this License issued under Section 6.1. For legal entities, You'' includes any entity which controls, is controlled by, or is under common control with You. For purposes of this definition, control'' means (a) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (b) ownership of more than fifty percent (50%) of the outstanding shares or beneficial ownership of such entity. -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3 -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
John Cowan wrote: Well, I could answer that in two, conflicting ways. If distribution becomes irrelevant, the spirit of the GPL in that respect is obsolete, isn't it? I don't see how that could happen, unless bandwidth (including the last mile) becomes too cheap to meter. I don't think that's relevant. Web Services will, by definition, be served this way. That they may not be available (or viable) to some is a different issue. Suppose A publishes a GPLed book describing some arcane subject, and B obtains a copy of it. C now mails questions to B along with payment, and B answers the questions out of the book and mails back the replies. In principle, C could read the book himself, but may not have the time or desire.) Surely A's rights are not impinged on here? Are things different if B adds his own marginal notes to the book? Is B really required by (the spirit of) the GPL to make those notes available to C? I see your point, but I still think that if we're going down that line you can say the same things about GPLd software (private mods installed at a client site, the client doesn't need the mods, so what's the problem, eh?). Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Russell Nelson wrote: their users the freedom I've given them, or even an acknowledgement of my work or a notice of where to get a copy of their own, because the service-users will never get to see the source and/or the license. And I'm not fine with that. That Way Lies Dragons. Acknowledged. Since the service provider is not copying the modified work when they let someone use it, the copyright holder (henceforth, 'you') needs more than copyright permissions to solve this problem. You need to get people to agree to distribute modifications[1], and you need to get them to agree to penalties if they refuse. Such an agreement would have to be enforcable in a court of law. As such, it would be a contract. A contract has to be negotiated. It cannot be applied unilaterally. The negotiation might be as simple as an offer and an acceptance. How do you distribute Open Source software in such a manner that the user can only access it after seeing an offer and performing an act which can only be interpreted as acceptance of that offer? [1] I don't see how such a license could be a free software license either, given what RMS has said about privacy in: http://www.gnu.org/philosophy/free-sw.html I see the issues this brings, but my immediate choices are 1) keep the source closed, or 2) release the sources under conditions that give the users of the service that my software will provide the same freedoms as the entity running that software for them has. I'm not up to speed on copyright law, but my choices here are (unfortunately) simple. If 2) can't be achieved, it's going to be 1) Emile -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
RE: OSD modification regarding what license can require of user
Mitchell Baker wrote: As to the question of the scope of a corporate entity, the MPL uses a control standard often found in corporate documents, securities documents and statutes. The language is the sort everyone hates -- clearly written by lawyers for laawyers. I've included it below. But it is in general use, is generally understood by lawyers and judges who may someday look at this, and so won't need to develop its own case law and understanding. There is a couple of other standards in use for determinging control but the one in the MPL seemed the best, at least when I researched this last. Mitchell *++ 1.12. You'' (or Your) * means an individual or a legal entity exercising rights under, and complying with all of the terms of, this License or a future version of this License issued under Section 6.1. For legal entities, You'' includes any entity which controls, is controlled by, or is under common control with You. For purposes of this definition, control'' means (a) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (b) ownership of more than fifty percent (50%) of the outstanding shares or beneficial ownership of such entity. I copied this section of the MPL in the Jabber license, with only slight changes, because the MPL got it right. Here's how the Jabber license says it: 14. Definition of You in This License. You throughout this License, whether in upper or lower case, means an individual or a legal entity exercising rights under, and complying with all of the terms of, this License or a future version of this License issued under Section 7. For legal entities, you includes any entity that controls, is controlled by, or is under common control with you. For purposes of this definition, control means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity. -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3
Re: OSD modification regarding what license can require of user
Mitchell, A possibly naive question: The text you submitted is a _broad_ definition that is in common use. Is there a similar _narrow_ definition as well? I don't see that this text would be the right way for a quid-pro-quo license to define the legal entity in which distribution doesn't happen, because that entity would include beta-testers under contract, would it not? Maybe even _users_ under contract or NDA? On the other hand, there are applications in a quid-pro-quo license _would_ use this definition, Your Licensed Patents comes to mind. Thanks Bruce -- license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3