Re: [License-discuss] plain text license versions?
Bruce Perens writes: >I question why you didn't call a halt when the discussion was >obviously becoming a testosterone contest past the point of any useful >content. OK, you'll never have the time to moderate. That's fine. What >isn't fine is that you don't find someone else to do it. Well, I've ever been able to moderate on a regular basis, and have done it even less since switching from Licensing Committee to Infrastructure Committee (within OSI, that is -- obviously, I'm still on this list and participating). Moderation is hard work, and no one's paying for it right now, so we'll have to rely on the group's collective good sense. In this case I shouldn't have stuck my hand in after the fact anyway -- it was an unnecessary provocation, and I probably caused more trouble than I saved! Sorry for the extra noise. -K pgp9GGKqwDtHq.pgp Description: PGP signature ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Quoting Bruce Perens ([email protected]): > I question why you didn't call a halt when the discussion was > obviously becoming a testosterone contest past the point of any > useful content. I'll tell you offlist why this is a hilarious characterisation. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On 09/10/2012 01:38 PM, Rick Moen wrote: Quoting Karl Fogel ([email protected]): It's better to question reasoning than motivations, on this list and probably most others. Karl, I question why you didn't call a halt when the discussion was obviously becoming a testosterone contest past the point of any useful content. OK, you'll never have the time to moderate. That's fine. What isn't fine is that you don't find someone else to do it. <> smime.p7s Description: S/MIME Cryptographic Signature ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Quoting Karl Fogel ([email protected]): > It's better to question reasoning than motivations, on this list and > probably most others. I said nothing whatsoever about motivations. Lack of either serious attitude, attention to the subject, and/or a sense of perspective seemed to exist as a fact on the ground, and I therefore exited as I wished to not waste more time. I'll be glad to stipulate that your friend's motivations were glorious, if you wish. Beyond that, your friend's personal merits are doubtless impressive but off my list of topics. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Rick Moen writes: >Ben, I was not giving you a very serious reply (but rather a dismissive >one), because frankly I don't think you are approaching this discussion >with a serious attitude, attention to the subject, and/or a sense of >perspective. Personal defenses might be inappropriate here, I don't know, but FWIW: I've worked with Ben Reser in a number of projects and he's always approached discussions with a serious attitude and attention to the subject. You may still disagree with his reasoning, but that's a different matter. Nothing I saw in this thread -- I admit I skimmed sometimes -- indicated a lack of seriousness or attention on his part. It's better to question reasoning than motivations, on this list and probably most others. Thanks, -K ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Fri, Sep 7, 2012 at 12:25 PM, Rick Moen wrote: > I exited, Bruce. > > You talk to the gentleman if you wish. I left. I'm done as well. Ben Tilly's response made sense to me. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Quoting Bruce Perens ([email protected]): > Is this really a serious discussion? I exited, Bruce. You talk to the gentleman if you wish. I left. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On 09/07/2012 11:24 AM, Rick Moen wrote: I don't think you are approaching this discussion with a serious attitude, attention to the subject, and/or a sense of perspective. Is this really a serious discussion? It sounds to me more like a contest of how many silly things some of us can get away with doing or advising our clients to do, in avoiding a requirement that is brain-dead simple and no sweat for anyone to fulfill. Some lawyers and IP specialists enjoy sophomoric discussions of legal theory that has little value in real life. I guess they can blow off steam here, at least until it gets /too /annoying. I hope they don't waste their clients time on such things. <> smime.p7s Description: S/MIME Cryptographic Signature ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Quoting Ben Reser ([email protected]): > Depending on the license that may not be so easy. [...] Ben, I was not giving you a very serious reply (but rather a dismissive one), because frankly I don't think you are approaching this discussion with a serious attitude, attention to the subject, and/or a sense of perspective. Thus, I basically gave your reply the brushoff and exited. I apologise to the mailing list for speaking unpleasantly in doing so, but my exiting from this dialogue you are attempting still seems well justified and I am doing so. Possibly, you will find other participants who see fit to take your points seriously. Good fortune to them and to you, sir. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Fri, Sep 7, 2012 at 10:58 AM, Ben Tilly wrote: > The location of the license text is not a provision of the license. > Some licenses, for instance the GPL, actually say that you have to > distribute the license along with the work. Others leave the matter > silent. Either way the license is an open source license. Right I realized that the location of the license text isn't a provision of the license. > You may argue that software with an uninterpretable license is not > really open. This is not a problem. Open source does not mean > copyleft. A lot of open source licenses allow people to incorporate > the software in proprietary products, and you don't even have to be > told it is there. And yes that is exactly what I meant. You're right I wasn't really considering the permissive licenses. I can see why not including the license is an important ability for them. > I have heard people who have distributed embedded software with GPLed > components disagree with this. Adding the GPL inside of a device that > nobody can interact with the inside of is pretty useless, and is > frustrating when they are often left fighting for every byte. Given > the number of devices with embedded computers, this is not exactly a > small use case. Another excellent point. Thanks for your insightful post. I can see why people might want to waive the right to require the license. I do still think as an author it's incredibly foolish and counter productive to not include the license text with your work. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Fri, Sep 7, 2012 at 10:37 AM, Ben Reser wrote: > On Thu, Sep 6, 2012 at 11:12 PM, Rick Moen wrote: [...] >>> For that matter is it not also a violation of the technology neutral >>> clause of the open source definition? >> >> No. Read it. > > I did. If you only provide a URL to read the license I can't fathom > how you're going to argue that doesn't violate that clause of the > definition. > > Which says: > "No provision of the license may be predicated on any individual > technology or style of interface." The location of the license text is not a provision of the license. Some licenses, for instance the GPL, actually say that you have to distribute the license along with the work. Others leave the matter silent. Either way the license is an open source license. You may argue that software with an uninterpretable license is not really open. This is not a problem. Open source does not mean copyleft. A lot of open source licenses allow people to incorporate the software in proprietary products, and you don't even have to be told it is there. > You can't even read the license without using a several specific > technologies. If it's not a common license like the GPL you may have > no way of knowing what the license said. It may not be in a book. > The URL may have disappeared. Making the software in essence All > Rights Reserved as far as anyone is concerned since they may no longer > have a clue what the license said. [...] This is all true, and entirely irrelevant to the point that you wish to argue. > All of these scenarios are easily overcome by just including the > license with the software. The licenses are not very large and I > don't see the problem with needing to include them. I have heard people who have distributed embedded software with GPLed components disagree with this. Adding the GPL inside of a device that nobody can interact with the inside of is pretty useless, and is frustrating when they are often left fighting for every byte. Given the number of devices with embedded computers, this is not exactly a small use case. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thu, Sep 6, 2012 at 11:12 PM, Rick Moen wrote: > Card catalogue? Help from a librarian? Depending on the license that may not be so easy. I doubt there is just a book of open source licenses. So you're likely going to have to find some other book that includes the license. If it's not a common license it may not even be possible. >> For that matter is it not also a violation of the technology neutral >> clause of the open source definition? > > No. Read it. I did. If you only provide a URL to read the license I can't fathom how you're going to argue that doesn't violate that clause of the definition. Which says: "No provision of the license may be predicated on any individual technology or style of interface." You can't even read the license without using a several specific technologies. If it's not a common license like the GPL you may have no way of knowing what the license said. It may not be in a book. The URL may have disappeared. Making the software in essence All Rights Reserved as far as anyone is concerned since they may no longer have a clue what the license said. >> How do you know that you've found the correct license text from a >> search engine. > > Indeed, Moriarty the Master Criminal might have put decoy pages up! > Let's call Interpol. Your response implies that the only way this would happen is if it was intentional. I think the unintentional mistake is far more likely. If you have the license included with the software there can be no dispute about what the license the author intended. If the author puts a link to a URL that then disappears and some distributor finds a different copy of the license that is different in some subtle way you may end up with a crummy outcome for everyone. All of these scenarios are easily overcome by just including the license with the software. The licenses are not very large and I don't see the problem with needing to include them. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Quoting Ben Reser ([email protected]): > On Thu, Sep 6, 2012 at 6:13 PM, Rick Moen wrote: > > Very sad. So, the barrio occupant in question might need to check out a > > book with the licence text. Life's imperfect. > > How exactly would you find said book? Card catalogue? Help from a librarian? This covnersation is one of those where the other person keeps saying 'Well, what about people who {can't do that|live in the Kalahari Desert|etc}?' I'm not that devoted to wasting time, so I'm bowing out now. > For that matter is it not also a violation of the technology neutral > clause of the open source definition? No. Read it. > How do you know that you've found the correct license text from a > search engine. Indeed, Moriarty the Master Criminal might have put decoy pages up! Let's call Interpol. > This whole line of thinking just seems contrary to the Open Source > Definition in my opinion. As we say in California, thank you for sharing. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thu, Sep 6, 2012 at 6:13 PM, Rick Moen wrote: > Very sad. So, the barrio occupant in question might need to check out a > book with the licence text. Life's imperfect. How exactly would you find said book? Are you not assuming that they have libraries and book store infrastructure, which in many cases is lacking if the location is lacking in Internet infrastructure as well. I'd almost argue that this is discrimination against persons that don't have access to the license, since if you can't read and understand the license then your only real choice is to not use the software since you can't know what you need to comply with. For that matter is it not also a violation of the technology neutral clause of the open source definition? We've specifically disallowed click wrap licenses since they don't work well with FTP or CDROM distribution. How can we consider only providing a URL to a license technology neutral when it require Internet access and the specific usage of a protocol that can be represented via a URL (presumably HTTP). > Ten years ago, the more commonly cited URL would have actually been > www.gnu.org/copyleft/gpl.html . Which today provides information about all > versions to date. A decade ago, it provided information only about v2, > v1 was already vanishingly rare and had seldom ever been used anyway. > > However, for the sake of discussion, let's suppose someone said in a > copyleft statement > > Copyright (C) 2002 George Tirebiter. > > This program is free software; you can redistribute it and/or > modify it under the terms of the GNU General Public License > as published by the Free Software Foundation; either version 2 > of the License, or (at your option) any later version. Full > text of GNU General Public License version 2 can be found at > http://www.gnu.org/licenses/gpl.txt . > > Post-2007, a user goes to http://www.gnu.org/licenses/gpl.txt and sees > GNU General Public License version 3. > > Rats! If only Internet users were capable of using a facility to find > things. If it were invented, one might call it a 'search engine'. > > (I readily concur with the implied point that it's a bad idea to swap > out a licence text for a later version at the same URL, however.) How do you know that you've found the correct license text from a search engine. I could easily take and put a license text which has a typo in it and a small typo could have very large implications for the meaning of the licenses. The obvious answer is that my buggy copy of the license shouldn't be the highest ranked result. But it's very hard to know what search terms will lead where in the future. In my opinion not providing a fully copy of the license in some human readable form (even if that form requires something like an HTML viewer) should be a no brainer. The licenses are almost always trivially short compared to the size of what they are licensing except for all but the most trivial things. This whole line of thinking just seems contrary to the Open Source Definition in my opinion. Not that I disagree with providing a repository of canonical and correct copies of the licenses in whatever format people find useful. But I don't think providing a URL link to the license is a very good idea. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
"Lawrence Rosen" writes: >I'd count that as another reason *not* to provide plain text license >files. I think it would be FAR more useful to have a simple license >statement in the source tree of each program that points to the >OFFICIAL version of that license on the OSI website. This also avoids >the duplication of text -- with potential transcription or legal >errors -- in many source code trees, and completely avoids the need to >actually read the licenses if one trusts OSI. We don't have a choice about whether plain text licenses are shipped in program source trees. That's going to happen. We *do* have some influence over whether those plain text versions are consistent, canonical, and correct. Therefore, etc, etc. I hope this makes my point clearer :-). -Karl ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Larry wrote: I think it would be FAR more useful to have a simple license statement in the source tree of each program that points to the OFFICIAL version of that license on the OSI website. You are very optimistic regarding the longevity of OSI. <> smime.p7s Description: S/MIME Cryptographic Signature ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Quoting Johnny Solbu ([email protected]): > On Thursday 06 September 2012 21:14, Lawrence Rosen wrote: > > I think it would be FAR more useful to have a simple license > > statement in the source tree of each program that points to the > > OFFICIAL version of that license on the OSI website. > > But it force the user to have internet access in order to actually > read the licence. There are many countries and places where internet > access is for the priviliged few. Very sad. So, the barrio occupant in question might need to check out a book with the licence text. Life's imperfect. > Besides, sometimes the license text changes. for example, 10 years ago > the license located at http://www.gnu.org/licenses/gpl.txt was GPLv2. > As of june 2007 it is GPLv3. Meaning that any gplv2 software that link > to the licese at that location is now linking to the Wrong license. Ten years ago, the more commonly cited URL would have actually been www.gnu.org/copyleft/gpl.html . Which today provides information about all versions to date. A decade ago, it provided information only about v2, v1 was already vanishingly rare and had seldom ever been used anyway. However, for the sake of discussion, let's suppose someone said in a copyleft statement Copyright (C) 2002 George Tirebiter. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. Full text of GNU General Public License version 2 can be found at http://www.gnu.org/licenses/gpl.txt . Post-2007, a user goes to http://www.gnu.org/licenses/gpl.txt and sees GNU General Public License version 3. Rats! If only Internet users were capable of using a facility to find things. If it were invented, one might call it a 'search engine'. (I readily concur with the implied point that it's a bad idea to swap out a licence text for a later version at the same URL, however.) ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thursday 06 September 2012 21:14, Lawrence Rosen wrote: > I think it would be FAR more useful to have a simple license statement in the > source tree of each program that points to the OFFICIAL version of that > license on the OSI website. But it force the user to have internet access in order to actually read the licence. There are many countries and places where internet access is for the priviliged few. Besides, sometimes the license text changes. for example, 10 years ago the license located at http://www.gnu.org/licenses/gpl.txt was GPLv2. As of june 2007 it is GPLv3. Meaning that any gplv2 software that link to the licese at that location is now linking to the Wrong license. Therefore at least one copy of the license text should always be included. One should Never depend on a license text beeing available online. The text can go away for whatever reason. > This also avoids the duplication of text On most distros they have a package containing common licenses, so one only have to package one copy of the license text, unless the license recuires that it be included. -- Johnny A. Solbu web site, http://www.solbu.net PGP key ID: 0xFA687324 Kom Arbeidslyst og treng deg på, her skal du motstand finne. signature.asc Description: This is a digitally signed message part. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thu, Sep 6, 2012 at 3:58 PM, Lawrence Rosen wrote: > That's unfortunate, because I advise it all the time for all licenses. > Anything more is a waste of time. And my clients have never been sued for > posting a link instead of a license. Maybe we are lucky??? The problem is not that you'll be sued for a technical violation. The problem is that when you are sued (or threatened with suit) for some more substantive violation, all your technical violations will be used to insinuate bad faith. Perhaps your clients have such spotless reputations and/or are so perfect in their substantive compliance that they aren't concerned about that, though :) Luis > /Larry (from my tablet and brief) > > Luis Villa wrote: > >>On Thu, Sep 6, 2012 at 2:37 PM, Lawrence Rosen wrote: >>> Is distribution of the *link* to the license sufficient compliance with >>> this requirement? >> >>For CC and MPL 2, yes. >> >>MIT and many others? The conventional interpretation is "no." >> >>Luis >> >>> /Larry (from my tablet and brief) >>> >>> Luis Villa wrote: >>> On Thu, Sep 6, 2012 at 12:14 PM, Lawrence Rosen wrote: > Karl Fogel wrote: >> Many coders expect to find plaintext license terms in a LICENSE or >> COPYING file, directly in the source tree. > > I'd count that as another reason *not* to provide plain text license > files. I think it would be FAR more useful to have a simple license > statement in the source tree of each program that points to the OFFICIAL > version of that license on the OSI website. This also avoids the > duplication of text -- with potential transcription or legal errors -- in > many source code trees, and completely avoids the need to actually read > the licenses if one trusts OSI. > > Doesn't CC do that, in a way, with their license logos? More specifically, CC does it with the requirement in the license that attribution notices link to the canonical text. Many OSS software licenses, unfortunately, require distribution of the actual text of the license. Luis ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thu, Sep 06, 2012 at 06:13:11PM -0400, John Cowan wrote: > Richard Fontana scripsit: > > > That assumes that the printed text is not "source code" in the sense > > meant in sections 1 and 2 of GPLv2 but is instead "object code or > > executable form" (section 3). I believe the better interpretation of > > GPLv2 is that text in a printed book is "source code ... in any > > medium" (the particular medium being printed text); thus you never > > reach section 3. > > "The source code for a work means the preferred form of the work for > making modifications to it." Leaving aside the fact that this definition only appears in section 3 (GPLv3 is different here), I regard the preferred form of human-readable text as human-readable text (e.g. not some encoded or obfuscated form), which GPLv2 section 1 says I can distribute "in any medium". "In any medium" implies that the source code (even if it is by definition "the preferred form for making modifications") can be distributed without triggering section 3 regardless of the nature of the medium of distribution. If your interpretation is correct then the question arises why the words "in any medium" appear in GPLv2 section 1. It doesn't say "You may copy and distribute verbatim copies of the Program's source code as you receive it, provided that you only do so in a machine-readable medium that is preferable as a medium for modification of the source code". IOW I think "form" and "medium" must mean different things in the GPL. - RF ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
That's unfortunate, because I advise it all the time for all licenses. Anything more is a waste of time. And my clients have never been sued for posting a link instead of a license. Maybe we are lucky??? /Larry (from my tablet and brief) Luis Villa wrote: >On Thu, Sep 6, 2012 at 2:37 PM, Lawrence Rosen wrote: >> Is distribution of the *link* to the license sufficient compliance with this >> requirement? > >For CC and MPL 2, yes. > >MIT and many others? The conventional interpretation is "no." > >Luis > >> /Larry (from my tablet and brief) >> >> Luis Villa wrote: >> >>>On Thu, Sep 6, 2012 at 12:14 PM, Lawrence Rosen wrote: Karl Fogel wrote: > Many coders expect to find plaintext license terms in a LICENSE or > COPYING file, directly in the source tree. I'd count that as another reason *not* to provide plain text license files. I think it would be FAR more useful to have a simple license statement in the source tree of each program that points to the OFFICIAL version of that license on the OSI website. This also avoids the duplication of text -- with potential transcription or legal errors -- in many source code trees, and completely avoids the need to actually read the licenses if one trusts OSI. Doesn't CC do that, in a way, with their license logos? >>> >>>More specifically, CC does it with the requirement in the license that >>>attribution notices link to the canonical text. Many OSS software >>>licenses, unfortunately, require distribution of the actual text of >>>the license. >>> >>>Luis ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thu, Sep 06, 2012 at 03:07:44PM -0700, Luis Villa wrote:
> As a practical matter, indicating, tracking and relying on waiver is a
> bit of a pain. e.g., lets say upstream says:
>
> "I give you a copy of the license this work is licensed under by
> pointing you at http://www.apache.org/licenses/LICENSE-2.0.html";
Not entirely unlike the standard ASF-recommended license notice in
specifying the canonical Apache License 2.0 URL.
> Downstream now has a problem: does this text constitute a waiver? Is
> this indication that we can "give any other recipients... a copy"
> (Apache 4.1) in the same manner? The easiest solution to this
> (admittedly small) problem is... to include a full copy of the
> license.
I have encountered situations similar to this and I typically treat
them as a waiver of the requirement to provide downstream recipients
with (in this case) a copy of the license text by doing more than
upstream bothered to do for me. I apply that principle so extensively
in open source contexts (not just to issues of license text inclusion)
that I ought to come up with some sort of Latin legal maxim for it.
> Or to put it another way: OSI spent a lot of time and energy
> discouraging people from using custom licenses. Custom waivers
> (particularly for something trivial like this) are just another form
> of the same mess.
The usual situation is that the waiver (if you accept my
characterization of it as a waiver) is implicit. The FSF has tried to
encourage the practice of explicit exceptions ("additional
permissions" in GPLv3 parlance) to literal (or presumed) license
requirements but it hasn't had much effect on behavior in the GPL
licensor community.
- RF
___
License-discuss mailing list
[email protected]
http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Quoting Luis Villa ([email protected]): > As a practical matter, indicating, tracking and relying on waiver is a > bit of a pain. e.g., lets say upstream says: > > "I give you a copy of the license this work is licensed under by > pointing you at http://www.apache.org/licenses/LICENSE-2.0.html"; The competent (and bog-standard) method of stating a waiver is inline in licensor's copyright notice. Copyright (C) 2012 George Tirebiter. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version, with the additional permission that licensor waives any requirement to include a copy of the license text. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. > Or to put it another way: OSI spent a lot of time and energy > discouraging people from using custom licenses. Custom waivers > (particularly for something trivial like this) are just another form > of the same mess. I cannot see that licensor voluntarily adding an extra right to the bundle otherwise conveyed creates a problem. However, whether OSI likes the practice or not, my point is that it occurs in the real world and solves practical problems. I don't have time to check implementation, but recall that this issue arose when my frined Marc Merlin created a custom package of the Exim MTA at VA Linux Systems that added OpenSSL integration for TLS/SMTPS and offered it for public download. I said to Marc 'Sorry to be the bearer of bad news, but your derivative of Phil Hazel's Exim code violates his copyright because OpenSSL includes GPL-incompatible code modules written by Eric A. Young back when it was called SSLeay.' Marc cursed me out (as if _I_ had caused the problem), and then did the obvious and asked Phil Hazel to add a licence exception permitting use with OpenSSL, which he reportedly did. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On 09/06/2012 03:07 PM, Luis Villa wrote: Custom waivers (particularly for something trivial like this) are just another form of the same mess. Posit that I am creating a version of the old Lyons Unix book, containing the Linux source code. How many copyright holders must grant me a waiver? Is the answer the same across all jurisdictions? It is easier to print the GPL than it is to even /start /analyzing questions like rights in a compilation vs. rights in a collective work. Thanks Bruce <> smime.p7s Description: S/MIME Cryptographic Signature ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Richard Fontana scripsit: > That assumes that the printed text is not "source code" in the sense > meant in sections 1 and 2 of GPLv2 but is instead "object code or > executable form" (section 3). I believe the better interpretation of > GPLv2 is that text in a printed book is "source code ... in any > medium" (the particular medium being printed text); thus you never > reach section 3. "The source code for a work means the preferred form of the work for making modifications to it." It was true at one time that people made modifications to books solely by writing in the margins and retyping, but that hasn't been true for a long time. -- Is not a patron, my Lord [Chesterfield],John Cowan one who looks with unconcern on a man http://www.ccil.org/~cowan struggling for life in the water, and when [email protected] he has reached ground encumbers him with help? --Samuel Johnson ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thu, Sep 6, 2012 at 2:55 PM, Rick Moen wrote: > Quoting John Cowan ([email protected]): > >> The difficulty is that text often winds up in printed books, and then >> you either have to distribute a CD with the book containing the editable >> source, or be prepared to issue such CDs for no more than the cost of >> distributing them. Both are expensive and awkward activities, and >> neither is well-supported by the printed-book sales channels that exist. > > Emphasis added: > > _Um, hello? Waiver._ As a practical matter, indicating, tracking and relying on waiver is a bit of a pain. e.g., lets say upstream says: "I give you a copy of the license this work is licensed under by pointing you at http://www.apache.org/licenses/LICENSE-2.0.html"; Downstream now has a problem: does this text constitute a waiver? Is this indication that we can "give any other recipients... a copy" (Apache 4.1) in the same manner? The easiest solution to this (admittedly small) problem is... to include a full copy of the license. Or to put it another way: OSI spent a lot of time and energy discouraging people from using custom licenses. Custom waivers (particularly for something trivial like this) are just another form of the same mess. Luis ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thu, Sep 06, 2012 at 05:45:00PM -0400, John Cowan wrote: > Rick Moen scripsit: > > > Years ago, I reminded readers on this mailing list that possibly useful > > reciprocal licences for non-software use by people disliking GFDL > > include GPLv2, and that FSF even published a piece explaining the > > advantages before they fell in love with GFDL: > > The difficulty is that text often winds up in printed books, and then > you either have to distribute a CD with the book containing the editable > source, or be prepared to issue such CDs for no more than the cost of > distributing them. Both are expensive and awkward activities, and > neither is well-supported by the printed-book sales channels that exist. That assumes that the printed text is not "source code" in the sense meant in sections 1 and 2 of GPLv2 but is instead "object code or executable form" (section 3). I believe the better interpretation of GPLv2 is that text in a printed book is "source code ... in any medium" (the particular medium being printed text); thus you never reach section 3. - RF ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Quoting Mike Linksvayer ([email protected]): > GFDL requires copy of license text. And you thought 'waiver' meant...? > Anyway, I like the option to refer to a license rather than include it That would be one sort of provision a waiver might state. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thu, Sep 6, 2012 at 2:33 PM, Rick Moen wrote: > Quoting Luis Villa ([email protected]): >> More specifically, CC does it with the requirement in the license that >> attribution notices link to the canonical text. Many OSS software >> licenses, unfortunately, require distribution of the actual text of >> the license. The actual text of the license is an option for CC licenses, but a link is almost always used instead (one or the other is required). > Not an objection, but just as a reminder: Licensor can waive that > requirement. > > Years ago, I reminded readers on this mailing list that possibly useful > reciprocal licences for non-software use by people disliking GFDL > include GPLv2, and that FSF even published a piece explaining the > advantages before they fell in love with GFDL: > https://www.gnu.org/philosophy/nonsoftware-copyleft.html > > I was told, here: You shouldn't do that. That's dumb, because then > redistributors would need to include the full text of GPL. > > Um, hello? Waiver. GFDL requires copy of license text. Anyway, I like the option to refer to a license rather than include it (CC licenses are specific about how to refer, MPL2 just says to state how to obtain a copy; I don't know which is better) and if license does not give this option, consider requirement to include a copy a not compelling reason to use an incompatible license. It is too bad "non-software" licenses constitute a largely separate universe to the extent it is for such reasons. And +1 to a repository of .txt renditions. Mike ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Quoting John Cowan ([email protected]): > The difficulty is that text often winds up in printed books, and then > you either have to distribute a CD with the book containing the editable > source, or be prepared to issue such CDs for no more than the cost of > distributing them. Both are expensive and awkward activities, and > neither is well-supported by the printed-book sales channels that exist. Emphasis added: _Um, hello? Waiver._ ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thu, Sep 6, 2012 at 2:37 PM, Lawrence Rosen wrote: > Is distribution of the *link* to the license sufficient compliance with this > requirement? For CC and MPL 2, yes. MIT and many others? The conventional interpretation is "no." Luis > /Larry (from my tablet and brief) > > Luis Villa wrote: > >>On Thu, Sep 6, 2012 at 12:14 PM, Lawrence Rosen wrote: >>> Karl Fogel wrote: Many coders expect to find plaintext license terms in a LICENSE or COPYING file, directly in the source tree. >>> >>> I'd count that as another reason *not* to provide plain text license files. >>> I think it would be FAR more useful to have a simple license statement in >>> the source tree of each program that points to the OFFICIAL version of that >>> license on the OSI website. This also avoids the duplication of text -- >>> with potential transcription or legal errors -- in many source code trees, >>> and completely avoids the need to actually read the licenses if one trusts >>> OSI. >>> >>> Doesn't CC do that, in a way, with their license logos? >> >>More specifically, CC does it with the requirement in the license that >>attribution notices link to the canonical text. Many OSS software >>licenses, unfortunately, require distribution of the actual text of >>the license. >> >>Luis ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thu, Sep 06, 2012 at 02:37:38PM -0700, Lawrence Rosen wrote: > Is distribution of the *link* to the license sufficient compliance with this > requirement? For licenses that appear literally to require inclusion of a copy of the license text? I have wondered whether we ought to start treating that as a reasonable modern interpretation of such requirements, given that many developers aren't bothering to bundle license texts to begin with. - RF > > /Larry (from my tablet and brief) > > Luis Villa wrote: > > >On Thu, Sep 6, 2012 at 12:14 PM, Lawrence Rosen wrote: > >> Karl Fogel wrote: > >>> Many coders expect to find plaintext license terms in a LICENSE or > >>> COPYING file, directly in the source tree. > >> > >> I'd count that as another reason *not* to provide plain text license > >> files. I think it would be FAR more useful to have a simple license > >> statement in the source tree of each program that points to the OFFICIAL > >> version of that license on the OSI website. This also avoids the > >> duplication of text -- with potential transcription or legal errors -- in > >> many source code trees, and completely avoids the need to actually read > >> the licenses if one trusts OSI. > >> > >> Doesn't CC do that, in a way, with their license logos? > > > >More specifically, CC does it with the requirement in the license that > >attribution notices link to the canonical text. Many OSS software > >licenses, unfortunately, require distribution of the actual text of > >the license. > > > >Luis > ___ > License-discuss mailing list > [email protected] > http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Rick Moen scripsit: > Years ago, I reminded readers on this mailing list that possibly useful > reciprocal licences for non-software use by people disliking GFDL > include GPLv2, and that FSF even published a piece explaining the > advantages before they fell in love with GFDL: The difficulty is that text often winds up in printed books, and then you either have to distribute a CD with the book containing the editable source, or be prepared to issue such CDs for no more than the cost of distributing them. Both are expensive and awkward activities, and neither is well-supported by the printed-book sales channels that exist. -- What is the sound of Perl? Is it not the John Cowan sound of a [Ww]all that people have stopped [email protected] banging their head against? --Larryhttp://www.ccil.org/~cowan ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Is distribution of the *link* to the license sufficient compliance with this requirement? /Larry (from my tablet and brief) Luis Villa wrote: >On Thu, Sep 6, 2012 at 12:14 PM, Lawrence Rosen wrote: >> Karl Fogel wrote: >>> Many coders expect to find plaintext license terms in a LICENSE or >>> COPYING file, directly in the source tree. >> >> I'd count that as another reason *not* to provide plain text license files. >> I think it would be FAR more useful to have a simple license statement in >> the source tree of each program that points to the OFFICIAL version of that >> license on the OSI website. This also avoids the duplication of text -- with >> potential transcription or legal errors -- in many source code trees, and >> completely avoids the need to actually read the licenses if one trusts OSI. >> >> Doesn't CC do that, in a way, with their license logos? > >More specifically, CC does it with the requirement in the license that >attribution notices link to the canonical text. Many OSS software >licenses, unfortunately, require distribution of the actual text of >the license. > >Luis ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Quoting Luis Villa ([email protected]): > More specifically, CC does it with the requirement in the license that > attribution notices link to the canonical text. Many OSS software > licenses, unfortunately, require distribution of the actual text of > the license. Not an objection, but just as a reminder: Licensor can waive that requirement. Years ago, I reminded readers on this mailing list that possibly useful reciprocal licences for non-software use by people disliking GFDL include GPLv2, and that FSF even published a piece explaining the advantages before they fell in love with GFDL: https://www.gnu.org/philosophy/nonsoftware-copyleft.html I was told, here: You shouldn't do that. That's dumb, because then redistributors would need to include the full text of GPL. Um, hello? Waiver. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Thu, Sep 6, 2012 at 12:14 PM, Lawrence Rosen wrote: > Karl Fogel wrote: >> Many coders expect to find plaintext license terms in a LICENSE or >> COPYING file, directly in the source tree. > > I'd count that as another reason *not* to provide plain text license files. I > think it would be FAR more useful to have a simple license statement in the > source tree of each program that points to the OFFICIAL version of that > license on the OSI website. This also avoids the duplication of text -- with > potential transcription or legal errors -- in many source code trees, and > completely avoids the need to actually read the licenses if one trusts OSI. > > Doesn't CC do that, in a way, with their license logos? More specifically, CC does it with the requirement in the license that attribution notices link to the canonical text. Many OSS software licenses, unfortunately, require distribution of the actual text of the license. Luis ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Karl Fogel wrote: > Many coders expect to find plaintext license terms in a LICENSE or > COPYING file, directly in the source tree. I'd count that as another reason *not* to provide plain text license files. I think it would be FAR more useful to have a simple license statement in the source tree of each program that points to the OFFICIAL version of that license on the OSI website. This also avoids the duplication of text -- with potential transcription or legal errors -- in many source code trees, and completely avoids the need to actually read the licenses if one trusts OSI. Doesn't CC do that, in a way, with their license logos? /Larry -Original Message- From: Karl Fogel [mailto:[email protected]] Sent: Wednesday, September 05, 2012 8:09 AM To: [email protected] Cc: [email protected] Subject: Re: [License-discuss] plain text license versions? "Lawrence Rosen" writes: >> Have we (OSI) ever seriously adding putting plain text versions of >> licenses (where available) to the OSI website? > >While this makes no difference to the legal implications of a license, >converting to plain text destroys information useful for human beings >to comprehend the license. It is like removing indentation and line >endings from source code. Please don't encourage old-fashioned ways of >representing licenses so they can't be easily read by the only ones >that matter: Human beings. > >This is part of my existential battle, including within Apache, to >acknowledge that HTML allows for a richer vocabulary of expression. >Quit down-versioning our creative works. :-) Does this qualify as a >"historical? >technical? inertial? other?" reason in your lexicon? Whichever, why >waste time creating an 80-column ASCII format in this day and age? Some >people, I guess, still use punched cards for their software, but let's >ignore their needs. Actually, I think we should provide plain text versions. (See http://projects.opensource.org/redmine/issues/8, which is about this.) Many coders expect to find plaintext license terms in a LICENSE or COPYING file, directly in the source tree. While they can of course still understand the text if it's in HTML, they prefer plain text -- and their editor software will often display HTML as raw markup rather than as a pretty page. So there's a very relevant group who want & expect plain text versions. When we don't provide those versions, those people sometimes manually reformat our HTML pages [1] in order to produce a plaintext file, occasionally introducing errors or at least inconsistencies. It's better if we just provide canonical versions. Karl [1] I have had to do this on more than one occasion. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Quoting Bruce Perens ([email protected]): > HTML provides some desirable features: Add the fact that it autoconverts very nicely to plaintext. ;-> $ grep html .mailcap text/html; /usr/bin/lynx -dump -force_html '%s'; copiousoutput; description=HTML Text; nametemplate=%s.html $ -- Cheers, "Overheard a hipster say 'Quinoa is kind of 2011', Rick Moenso I lit his beard on fire." -- Kelly Oxford [email protected] McQ! (4x80) ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Arguing the merit of plain text vs. HTML is just Lilliput v. Blefuscu. Provide both, for different reasons. Plain-text is a better source for cut-and-paste operations. In general plain text divides the actual license text from any attached commentary, making it clear which is which. There is no universally-accepted standard for indicating the character set of plain text in the data, rather than in an external indication such as the HTTP Content-Type header. There is an assumption, sometimes wrong, that plain-text is ASCII. ASCII isn't capable of representing non-Latin character sets. Web servers often misrepresent the character set of plain text, since the content-type indication is set in an external file rather than the content itself. HTML provides some desirable features: Web page producers are more conscious of the need to represent the page character set accurately. It is possible for a web site to enforce that all pages be UTF-8, and for most national characters to be represented accurately. However, not all sites are this well-disciplined, and there are regional issues such as the Han unification in UTF-8 that can cause an undesirable rendering of a character for languages like Japanese. In logogramic languages, getting a character wrong in a legal document is much more likely to cause an unintended change of meaning. This is not to say that plain text could render these characters at all. HTML provides internal anchors which can be referenced by external documents, providing a way to link to a particular paragraph (or finer, if provided) from an email or article. HTML provides a wealth of methods for rendering commentary internal to the document. It can be called out by changes in font, color, or position. It can be hidden and revealed using javascript, CSS, or document structure, and selected by hover-over or click. <> smime.p7s Description: S/MIME Cryptographic Signature ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
On Wed, Sep 5, 2012 at 8:08 AM, Karl Fogel wrote:
> "Lawrence Rosen" writes:
>>> Have we (OSI) ever seriously adding putting plain text versions of
>>> licenses (where available) to the OSI website?
>>
>>While this makes no difference to the legal implications of a license,
>>converting to plain text destroys information useful for human beings to
>>comprehend the license. It is like removing indentation and line endings
>>from source code. Please don't encourage old-fashioned ways of representing
>>licenses so they can't be easily read by the only ones that matter: Human
>>beings.
>>
>>This is part of my existential battle, including within Apache, to
>>acknowledge that HTML allows for a richer vocabulary of expression. Quit
>>down-versioning our creative works. :-)
What richer vocabulary did you have in mind? I'm open to the idea, but
my experience in creating MPL (and doing some very amateur study of
typography in legal documents) is that most formatting of legal
documents is either (1) fairly easy to replicate meaningfully in plain
text or (2) not very good typography.
For example, compare:
https://www.mozilla.org/MPL/2.0/ (typographically sophisticated HTML
license, using cutting-edge browser features)
https://www.mozilla.org/MPL/2.0/index.txt ("plain" ASCII text,
generated by pandoc from Markdown)
The HTML version is literally cutting-edge, both in the sense of
attempting to comply with best practices for legal typography, and in
using some CSS features that were not available in a stable Firefox
release until slightly after the license was released :)
Despite that, the plain text version is (1) not lossy (very
important!) and (2) looks pretty good.
Some things that are "lost" in the transition:
* use of Small Caps for titles. This makes things look less beautiful,
but no information is lost.
* use of typographically-correct quotes. Ditto - less beautiful, no
information lost.
* hyphenation/justification. Ditto.
* fonts. Ditto.
The one thing I can think of offhand that is slightly messy is
bold/italic, but those should really be used less in legal documents,
not more (see, e.g., discussion at
http://www.typographyforlawyers.com/?page_id=1770 )
> Actually, I think we should provide plain text versions. (See
> http://projects.opensource.org/redmine/issues/8, which is about this.)
I'm very uncomfortable providing text versions where the author hasn't
provided a canonical version.
> Many coders expect to find plaintext license terms in a LICENSE or
> COPYING file, directly in the source tree. While they can of course
> still understand the text if it's in HTML, they prefer plain text -- and
> their editor software will often display HTML as raw markup rather than
> as a pretty page.
Note also that asking the various license parsing tools out there to
grok LICENSE.html rather than just LICENSE.txt would be a pain. In
general, the licensing community is trying to make licensing
information *more* amenable to being automatically parsed, not less,
and I think OSI should definitely support that.
Luis
___
License-discuss mailing list
[email protected]
http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
"Lawrence Rosen" writes: >> Have we (OSI) ever seriously adding putting plain text versions of >> licenses (where available) to the OSI website? > >While this makes no difference to the legal implications of a license, >converting to plain text destroys information useful for human beings to >comprehend the license. It is like removing indentation and line endings >from source code. Please don't encourage old-fashioned ways of representing >licenses so they can't be easily read by the only ones that matter: Human >beings. > >This is part of my existential battle, including within Apache, to >acknowledge that HTML allows for a richer vocabulary of expression. Quit >down-versioning our creative works. :-) Does this qualify as a "historical? >technical? inertial? other?" reason in your lexicon? Whichever, why waste >time creating an 80-column ASCII format in this day and age? Some people, I >guess, still use punched cards for their software, but let's ignore their >needs. Actually, I think we should provide plain text versions. (See http://projects.opensource.org/redmine/issues/8, which is about this.) Many coders expect to find plaintext license terms in a LICENSE or COPYING file, directly in the source tree. While they can of course still understand the text if it's in HTML, they prefer plain text -- and their editor software will often display HTML as raw markup rather than as a pretty page. So there's a very relevant group who want & expect plain text versions. When we don't provide those versions, those people sometimes manually reformat our HTML pages [1] in order to produce a plaintext file, occasionally introducing errors or at least inconsistencies. It's better if we just provide canonical versions. Karl [1] I have had to do this on more than one occasion. ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
Re: [License-discuss] plain text license versions?
Luis Villa asked: > Have we (OSI) ever seriously adding putting plain text versions of > licenses (where available) to the OSI website? While this makes no difference to the legal implications of a license, converting to plain text destroys information useful for human beings to comprehend the license. It is like removing indentation and line endings from source code. Please don't encourage old-fashioned ways of representing licenses so they can't be easily read by the only ones that matter: Human beings. This is part of my existential battle, including within Apache, to acknowledge that HTML allows for a richer vocabulary of expression. Quit down-versioning our creative works. :-) Does this qualify as a "historical? technical? inertial? other?" reason in your lexicon? Whichever, why waste time creating an 80-column ASCII format in this day and age? Some people, I guess, still use punched cards for their software, but let's ignore their needs. My opinionated best regards, /Larry -Original Message- From: Luis Villa [mailto:[email protected]] Sent: Tuesday, September 04, 2012 9:19 AM To: License Discuss Subject: [License-discuss] plain text license versions? Have we (OSI) ever seriously adding putting plain text versions of licenses (where available) to the OSI website? Just off the top of my head, Mozilla, FSF, and Apache all provide both HTML and plain text versions of their licenses, and I'm sure others must as well. Seems like it would be useful for OSI to do this as well. That said, it hasn't already been done, I'm guessing there is some reason (historical? technical? inertial? other?) for it. Anyone know what that reason is? Luis ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss ___ License-discuss mailing list [email protected] http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
