On Thu, Jun 13, 2002 at 12:40:19PM -0700, Walter Landry wrote: > Branden Robinson <branden@debian.org> wrote: > > 5) There will be text, in the form of a brief notice, following the > > copyright notice, which mentions endorsements. Removing that text > > will not be permitted, unfortunately (yes, this is invariant text). > > You realize, of course, that this will make DFCL stuff unusable with > GPL'd code. Do you really want to do that? As noted elsewhere, I'm planning on a "GPL conversion clause". This would permit the omission of the endorsements notice. Actually, it would only "suspend" it. The idea being that you can't take a DFCL-licensed work, cram it into a piece of software, and then suck it back out into a data file without having to restore the Endorsements section. However, as long as the DFCLed work "lived" within a GPLed work, reproduction of the endorsement clause would not be required (remember, the endorsements themselves are *never* required). I don't see any GPL-compatibility problems with that. Contrary to popular belief, the GPL does not change the copyright on anything. It merely requires that you not impose restrictions beyond its own terms. A DFCL-licensed work is independently copyrighted and licensed. If you incorporate that work into a GPLed one, the endorsement terms would be "masked off", but would re-assert themselves once the independently-copyrighted were were extracted from its GPLed container. > I feel that endorsements should be separate from the work. Something > along the lines of "This manual may have changed since I wrote it. I > endorse the FooBar Manual with MD5 hash ... On most systems, you can > check the MD5 hash of this document by doing XYZ." The author could > even attach it to the document. It's worse than useless to have an MD5 hash on a printed book -- even if it's not mandatory to print it, but especially if it is. > If someone changes the document, then the hash will not match. If > they change the corresponding MD5 hash in the endorsement, then you've > got fraud. Let me try and offer a little more insight into my thinking. Fundamentally, I think both of the "advertising clauses" in the traditional BSD license are inappropriate for a copyright license. The reason I believe this is because a name or phrase is not something you can copyright (though some people have tried[1]). Copyrights apply to specific expressions of ideas. Copyrights do not apply to ideas in general (that's the realm of patent law), and they not applied to simple labels or names for things (that's the realm of trademark law). If a thing short enough to trademark, it is far too simple to merit copyright protection. In my opinion, a copyright license should not confuse the issue. The issue is that it's a copyright license. A copyright license is not the place to talk about patents, trademarks, libel, defamation of character, murder, treason, zoning laws, estate taxes, or collateral estoppel. Any copyright license that muddies the waters of copyright with anything *but* copyright is thus, to some extent in my view, bullshit. Now, before anyone gets carried away, I'm not about to advocate that the DFSG be amended such that any license with "bullshit" in it is excluded. My idealist heart yearns for the day, but there are too many practical problems at present, and many free software licenses have to dirty their hands with "bullshit" because proprieteers have already muddied the waters, and because judges and legislators have made bad decisions, letting copyright holders get away with murder. The point of my tangent is this: we don't *need* statements like "Neither the name of the University nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission." in our licenses. Why? For the same reasons I can't get a job as an editor at the newspaper and run a story saying "the Board of Regents of the University of California issued a statement Thursday saying that Linux is, like, K-Rad, and, ohmigod, the Free Software Foundation is, like, the coolest thing eh-ver! And what about Microsoft? *GAG ME* with a *SPOON*!" If I try to pull something like that, I'll be hearing from UC's lawyers, and they won't be talking to me about copyright. Similarly, I can't buy advertising time on cable and run an aid for mail-order peppermints in the shape of penguins, with a voice-over that says "Linus Torvalds says they're 'LINUX-LICIOUS'!". If I do, I'll be hearing from Linus's lawyers, and they won't be talking to me about copyright. They'll be talking to me about the Linux trademark, and they'll be talking to me about Linus's name. And guess what? Linus's name is not trademarked (as far as I know -- Billy Joel's is, though), copyrighted, or patented. Nevertheless he *still* has a civil cause of action against me. And that's true in just about every country in the world. So, with all that said, why am I in favor of an Endorsements section? Because it seems to be something that authors want. I can also sympathize with their desire for clarity as to whether or not they want to be associated with a given work or not. If an author creates a work and places it in the commons by licensing it under the DFCL, she has generously given us rights to her work. What she *hasn't* given us rights to is her name. Because extremely proprietary approaches to copyright have been the norm for so long, people likely *are* going to assume that when they pick up something that says "Copyright (C) Foo Bar" that the work was released with the approval of Foo Bar, and that Foo Bar signed off on every word in the work. With any copyleft -- not to mention a BSD/MIT-style copyright -- that assumption is going to be unwarranted. So what I am trying to achieve is a means of countering it. And I think one good way to counter it would be to have an endorsements notice right under the copyright notice. We *could* go a route that only precludes and disavows endorsements, e.g., Copyright (C) 200 Foo Bar This work may be freely copied, modified and distributed under the terms of the Debian Free Content License. The contents of this work are not necessarily the responsibility of the copyright holder(s) and do not necessarily represent the views of the copyright holder(s). In fact, I won't have any fundamental objection if we go this route with the DFCL. *However*, I think we could make the DFCL more appealing to authors if we did actually permit the inclusion of endorsements. These would consists of names only. No dates, no conditions, no digital signatures, no checksums, no commentary. A simple endorsement, on or off. Not only is this simpler and more space-efficient, but it is completely unambiguous as regards copyright. A name is not copyrightable. We don't run into any muddy water with people making "endorsement statements" and insisting that these be invariant. It boils down to a simple question: Do you want your name listed as endorsing this version or not? I like that simplicity. Why do endorsements at all? As I said before, I think this would make the DFCL more appealing to authors, and to interested third parties as well. Think of works that hew more towards the artistic. If I write my Great American Novel but cannot find a publisher, I can DFCL it, and at least I can affix my endorsement to the version that reflects my artistic integrity. Endorsements would be good for standards documents as well. You could have a DFCL-licensed RFC, and the ISOC would probably get what they want; unambiguous notice that they are the copyright holders of the document, but also explicit notice that derived works don't meet with their approval unless they say so. Then they might stop being tempted to violate DFSG 6. See <http://www.faqs.org/rfcs/rfc-editor/rfc-copyright-story.html>. [1] http://216.239.37.100/search?q=cache:bzDmKFqDSnYC:www.securedparty.org/news/stlouis.htm+defendant+%22claims+copyright+on+his+name%22&hl=en&ie=UTF8 -- G. Branden Robinson | "I came, I saw, she conquered." Debian GNU/Linux | The original Latin seems to have branden@debian.org | been garbled. http://people.debian.org/~branden/ | -- Robert Heinlein
Attachment:
pgpu1usrNHZAU.pgp
Description: PGP signature