Subject:
Re: LDP Licence at http://tldp.org/COPYRIGHT.html, and changing in-place
From:
Rick Moen ####@####.####
Date:
27 Dec 2003 05:46:25 -0000
Message-Id: <20031227054624.GA1166@linuxmafia.com>
Quoting David Lawyer ####@####.####
We need to stop this, David. It's going _nowhere_. One last effort:
> Sorry, it's like a typo. I typed it correctly but I have a macro "gregf"
> that changed what I typed and then I removed that but never got the
> "gregf" back. It should be:
>
> www.ibiblio.org/gregf/ldp_old
I get 404 _there_, too, with or without a trailing slash. Ditto searching
Wayback machine for those.
> I thought that one was v2.0 of LDPL and the other was the Copying
> License.
Sorry, my mistake. I realised that shortly after replying to you, and
accordingly corrected the filenames and descriptions at
http://linuxmafia.com/kb/Licensing_and_Law on the three entries, there.
Anyway, I have located a very early and unnumbered version of
> the LDPL. It's at
> http://www.ibiblio.org/pub/historic-linux/ftp-archives/sunsite.unc.edu/Nov-06-1994/docs/HOWTO/COPYRIGHT
Interesting. The wording has similarities to that of both the 16
September 1999 LDP Copying License and the 12 January 1998 LDPL v2.0
text.
But, correct me if I'm wrong, but licenses of interest are those
referenced by current LDP documentation as applying to them (or possibly
applying to them in cases of ambiguity). Is there any reason to believe
that's true of this licence? As before, you don't say, thereby leaving
(once again) the key factual basis for your claim unstated.
> From the wording, it's obvious that this is the predecessor of LDPL.
> But it's still not of too much help since it's too old.
To recap: What makes a licence relevant is its use (or possible use) by
one or more creative work of interest. My immediate interest is limited
to current LDP documents; others might care about licensing of
discontinued documents.
> I looked at most of the HOWTOs for 1994 (LDP got started in Jan. 1993)
> and found no HOWTOs that used it even though it was in the Manifesto.
So, you're saying that no HOWTOs in 1994 specified the text at
http://www.ibiblio.org/pub/historic-linux/ftp-archives/sunsite.unc.edu/Nov-06-1994/docs/HOWTO/COPYRIGHT
in any way as the applicable permission grant? OK. But I don't see the
relevance to present-day concerns.
> The statement that the Copying License applied to all LDP documents
> was a joke.
More precisely, it's a legal no-op. To recap: Control of licensing
terms covering any instance of a creative work rests entirely with its
copyright holder. The licence can "claim" all it wants about things its
author thinks it covers, but the licence is not a person and cannot own
property. Therefore, any licence language purporting to specify what it
covers has, in itself, no legal effect.
> So what this shows the that the Copying License was never of much
> significance in LDP.
Your use of the term "the Copying Licence", here, is ambiguous. You
could mean the text you found at
http://www.ibiblio.org/pub/historic-linux/ftp-archives/sunsite.unc.edu/Nov-06-1994/docs/HOWTO/COPYRIGHT,
the 6 January 1997 text I now have mirrored at
http://linuxmafia.com/faq/Licensing_and_Law/ldpc-license-1997-01-06.html,
the 16 September 1999 text I now have mirrored at
http://linuxmafia.com/faq/Licensing_and_Law/ldpc-license-1999-09-16.html,
or something else entirely.
In any event, I would be inclined to believe that of the middle one only
after examining substantively all LDP documents, given that it was
posted at http://www.tldp.org/COPYRIGHT.html , and thus likely linked
to.
> Before I found this out, I thought that perhaps in
> the early days of LDP the Copying License was widely used. Nope. In
> fact most HOWTOs in 1994 had their own copyright and many had no
> copyright at all and thus, according to the claim to the LDP license,
> fell under it's jurisdiction.
1. ARGH! Please do not use the term "copyright" to refer to a
permission grant (licence). That's _not_ a copyright, and the usage has
a long and sad record of confusing people. Copyright is an abstract
legal property right over most aspects of creative works instantiated in
fixed form (in some but not all categories of creative effort). That
right arises inherently in the act of creation and becomes the creator's
property (in Berne Convention countries). It can be conveyed in writing
to new owners, if the creator so wishes (likewise, Berne Convention).
2. A licence has, in itself, no power to extend its coverage over
anything at all. What a licence claims is under its jurisdiction
has no legal effect.
3. It is incorrect to state that some HOWTOs in 1994 had no copyright.
Of course they had copyrights, which arose automatically in the act of
creation. The instances of those HOWTOs uploaded to LDP in 1994 had
whatever permission grants attached to them that the copyright holders
expressed or implied through the circumstances of the time.
Let's suppose that I submitted a HOWTO in 1994. Some years later, I sue
someone using substantial passages from my downloaded HOWTO in a
derivative work (say, a book), claiming copyright infringement.
Defendent seeks to prove that he had permission for his usage.
Case 1: My HOWTO says nothing about copyright or licensing, but
submission history shows that I submitted it. Court finds that I own
copyright title. Defendent must show the court reason to believe that I
somewhere implied various permission grants above and beyond the implied
right to download my work from the LDP collection to which I submitted
it.
Case 2: My HOWTO asserts my authorship (and thus copyright), and says
something about a set of permissions somewhere that cover that instance
of my work. Court will seek to apply and interpret those permissions,
whatever they are.
In no case is some statement posted at the LDP going to automatically
apply merely because it's there and has delusions of grandeur. Such a
statement has, in itself, no power to apply to my property or anyone
else's.
> The Copying License also claimed to cover it but was not in the 1994
> snapshot of the LDP site.
This is likewise of no legal effect, for the same reason.
> The problem is that thus far, we haven't found any of these variants yet.
I'm sorry, you'll have to see my prior explanation of why the LUG HOWTO
and probably many others purport to have been covered by what is at
http://www.tldp.org/COPYRIGHT.html . That HOWTO was started by Kendall
Clark in early 1997. I added to it, as a derivative work of Kendall's
from 2002 to the present, during which time it purported to be covered
by licence text at that URL.
Now, as copyright holder over my portion of the work, I have sole power
to issue new instances of that work (my portion) under the terms of my
choosing -- but I don't have that power over Clark's earlier work
(unless Clark writes to me assigning me copyright ownership).
And you are once again ignoring my explanations of why my
recommendations hold regardless.
> I don't think anyone in LDP actually used v2.0.
Well, it looks like *I* did, for one.
> However, searching the Web, I find that 60 sites have copied v2.0 and
> many non-LDP docs use it as their license.
That would be one of the many reasons LDP should disambiguate the current
situation, to the extent possible, regardless of the questions you raised.
> Seems like a good idea but we haven't found any versions yet that people
> might have used.
That appears to be incorrect. And you're still ignoring my explanations
of why my recommendations hold regardless.
> Another problem is that the v2.0 has been used by others outside of
> LDP and we shouldn't break this link.
I wasn't proposing that LDP "break this link". To review, I suggested
having http://www.tldp.org/COPYRIGHT.html become an informational page
that attempts to disambiguate the matter, listing licences known to have
been at that URL in the past, with their dates of posting and
descriptions -- and links directly to their text, including that of
LDPL v2.0. Recommended licences, if any, could be thus indicated.
Deprecated licences could be thus indicated.
As authors notice -- courtesy of that informational page -- the
ambiguity that's been papered over until now, they would be able to take
corrective action as appropriate.
We've been over this before. Why are we having to go over it again?
> What we could do is try to find the missing versions and then email
> all LDP authors that link to COPYRIGHT.html with the info. Can
> someone search the ibiblio site for contents, looking for files that
> have no links to them? They might be in there somewhere.
That sounds like a whole lot of effort for no particular gain.
I've already handed you (plural) a suggestion with all the real work
done, on a platter, tied up with a nice bow. Either do it or don't.
But these irrelevant digressions are a waste of time.
> If someone used that license, would they not link to the url shown in
> that license?
One more time, David: If I write a HOWTO and specify what terms cover
it, I'm going to do so via (e.g.) a URL at which I see that licence.
I'm not bloody likely to snip a broken URL out of the HTML source of an
appendix in that licence -- especially if that broken URL isn't even
visible in the displayed Web page.
If you're saying someone actually _did_ that very peculiar and
staggeringly unlikely thing you speak of, show me. Where?
> And why do you state that is is an incorrect URL? The
> version 2.0 was likely put at this url.
Look, it's entirely irrelevant where it was, above and beyond the
locations at which it was actually linked to by LDP documents. If you
know of such locations, please show us the LDP documents.
> No. The words used to describe the license show the intent of the
> author.
Not when they're not the author's words, and not in his creation. Look,
it's the creative work and what it says about permissions that are
relevant. You keep trying to claim that some licence document up on the
Web has magic powers over copyright. Wrong. Licences cannot own
property. Licences cannot institute permssions grants. Copyright
owners have sole power to dictate licensing.
> This is the license they intended to use.
Not unless the _copyright owner_ so indicated. Once again, you post an
unlikely claim and fail to support it with an example. Show us.
> If the url they put into the license doesn't point to the license they
> intended to use (or the contents of this link altered), then that link
> was a mistake and should be corrected.
The link was a mistake solely because it's a mistake. The copyright
owners have nothing to do with it.
> I think that the Copying License put at COPYRIGHT.html around 1998 was
> put there by mistake.
Who cares? That's completely irrelevant to the question. What matters
is not what LDP intended, but rather what LDP authors specified using
their copyright powers. At least one such author, probably many,
specified that the terms at http://www.tldp.org/COPYRIGHT.html would
apply. That makes what was at that URL significant.
> Here's one scenario of how this might have happened. Someone notices
> that the LDP Licence (LDPL) at COPYRIGHT.html was not a copyright but
> a license, and moved it to LICENSE.html.
What do you mean by "was not a copyright"? I think you must be still
very, very confused about fundamental aspects of copyright law. Thus
the extremely fruitless recapitulation of increasingly large numbers of
basic points. Enough, please.
> Now it didn't apply to all works since people were using LDPL and a
> variety of other licenses.
No, that's not the reason it didn't apply to all works. It didn't apply
to all works because _copyright owners_ didn't apply it to all works.
David, you would not be still making this mistake if you had heeded my
clarification, many messages ago, that copyright owners have sole power
to specify terms under their copyright authority. Please: Stop.
Think.
> I think that for the past several years, what was at COPYRIGHT.html was
> not what the author intended.
What "author"? For purposes of the present discussion, the intentions
of the licence drafter are utterly insignificant. Only the expressed
will of the LDP docuements' copyright owners particularly matters.
> Thus one needs to find the 1.x versions of LDPL.
1. The conclusion is non-sequitur.
2. Nonetheless, you're welcome to go find some additional versions of
LDPL. Nobody's standing in your way. When you do, be sure to
state reasons to believe that the copyright owners of LDP documents
-- preferably current ones -- in some way specified that licence
text as applicable (or may have done so).
> > Where may I view the latter licence?
>
> All 1.x versions are lost (unless the one I just found might be called
> 1.0). So they remain to be found.
So, how is it we know that "LDPL v2.0 is somewhat close to version 1.x",
given that (once again!) _you didn't specify_?
> Well, the statement I made meant at the time I made it the v2.0 was the
> closest known version to what the HOWTOs intended to link to. In other
> words v2.0 was something like the 1.x versions that are lost. Now that
> I've found an unnumbered version, it may be that it's more similar to
> the 1.x versions that people linked to.
Well, for heaven's sake, post the whole lot of what's been found as
links from a new page at http://www.tldp.org/COPYRIGHT.html, with
information about what URLs they were at, posted on what dates, and
other things known about them.
> I meant: Add a note at the top of v2.0 stating that docs that linked to
> it prior to Jan, 1998 have linked to an earlier version. Such earlier
> versions are all lost. But in view of the fact that it's being used
> outside of LDP, perhaps we shouldn't add such a note. I'm not sure.
I would think that notes on the new http://www.tldp.org/COPYRIGHT.html
stating where and when the various linked licences were posted (and
which if any are now deprecated or recommended) would be the right
thing.
> I've clearly stated the reasons why I think that it's unlikely that any
> LDP author used these licenses via a link to COPYRIGHT.html. Now the
> obsolete Danish-HOWTO did use the Copying License, but it was done by
> linking to LDP-COPYRIGHT.
ARGH! You keep doing this! You make reference to something without
stating where it is, so that people can see and examine it.
> v2.0 was put there after it was deprecated and links to it from LDP
> removed. The fact that no one used the link specified in v2.0 also
> indicated that it wasn't used by anyone in LDP.
One more time: Look at the example of my Linux User Group HOWTO.
> The other one you found there is the Copying License. Yet most of the
> HOWTOs say LDP License (none say Copying license)....
Irrelevant, for reasons already mentioned.
> ...and I believe that the Copying License was put there by mistake.
Irrelevant, for reasons already mentioned.
> Well, LDPL was something like this. It claimed to apply to all docs
> with no license of their own (a default license).
That is, in itself, a legal no-op. But please note that (once again)
you've left it ambiguous what specific licence text you're referring to
when you say "LDPL was...."
> If someone gave a doc to the LDP, then, if they were told that if they
> didn't include a license it would come under the default license, then
> the default license is valid.
If they were told this, then they were told something that is factually
incorrect. The situation is ruled by copyright law: A judge (in a
copyright infringement suit) would examine the surrounding situation for
whatever best indicates the extent of permission granted by the
copyright holder through oral and written grants and through conduct of
parties in question.
The reason people create licences is for the copyright holder to
unambiguously indicate a non-default grant of otherwise reserved rights,
in such a way as to make it unnecessary for a judge to interpret those
surrounding circumstances.
> I don't know if they were told this or not, and if there is no
> meeting of minds (and thus no agreement reached) then for LDPL it's a
> no-op like you state).
ARGH! "Meeting of the minds" is a term in _contract_ law. This is not a
contract. This is copyright law, which is unilateral.
> But if you put in your howto certain permissions for someone else's
> howto, what does that mean?
If you don't own copyright over that second HOWTO, then your purported
action is a legal no-op. Moreover, the attempt to assert a reserved
right over someone else's property could in many jurisdictions be
considered, in itself, to constitute copyright violation -- if the
matter were ever sued over.
> You have no right to do this but that's what the Copying License did.
I already said: It's a legal no-op.
> > You are obsessing over meaningless trivia. That clause is likewise a
> > no-op, for the exact same reasons.
>
> My point is that I don't like the idea of linking to such a license.
If the consensus of LDP is that it's a badly drafted licence, put a note
in bold red type at the top saying "This is a badly drafted licence."
Frankly, I've seen a lot worse.
> I did this some time ago. A copyright owner is allowed to give others
> permission to do things with the copyrighted work.
That is correct. For example, I specify in my WordPerfect for Linux FAQ
that the recipient may elect to receive it either under GPLv2 or the
Creative Commons Attribution-ShareAlike 1.0 licence.
> One can permit someone else to make a derivative work or even permit
> one to change the license for the derivative work.
No, this is per statute a monopoly of the copyright holder. I suppose I
_could_ put up at the top of my WordPerfect for Linux FAQ a statement
that says "You may receive a copy of this work under any licence
whatsoever", but that's still me, the copyright holder, specifying
rights. I would speculate that a judge (in any related copyright dispute)
would probably rule, for reasons of sanity of legal proceedings, that I
had in effect issued a non-revocable licence to use the work for any
purpose without conditions -- because the law tends not to like having
people play "Nomic" with it, and an annoyed judge would be likely to
bring that high-flying concept back down to earth.
> Note that this doesn't change the license of the original since only
> the copyright owner can change that license.
I fear that is not quite correct about what happens in situations of
derivative works. The derivative must consist of elements with
_compatible_ licensing, else the original author's copyright will almost
certainly be violated.
For example, when a new-BSD-licensed work gets GPL-covered code added to
it, the resulting derivative isn't GPLed. The new code is GPL; the old
code is BSD. Each is an expression of the respective copyright owner's
permission grant, as to his property.
In that case, it isn't accurate to state that the original author
permitted the second one to "change the licence for the derivative
work". It's more accurate to state that the original author used a
licence (for his property) that has wide compatibility with subsequent
authors' choice of licensing for their additions, i.e., the new-BSD one.
In other cases, such as the Linux kernel's AIC-7770 driver, the
copyright holders have dual-licensed, i.e., you may receive the copy of
their property under either the provisions of new-BSD or GPLv2 licences,
at the recipient's option. This is still the original copyright owners
speaking, not the recipient. The recipient has no say whatsoever in the
scope of what is allowed.
> An interesting case arises if the copyright owner permits derivative
> works but fails to state what license they may be licensed under.
That's not particularly "interesting" -- or rare; it happens every day
someone receives a BSD-licensed or MIT/X-licensed piece of code, which
is to say, all the time.
People do this because they happen to have minimal or zero interest in
what licence people happen to choose for subsequent third-party additions.
> This seems to, unfortunately, be the case for Creative Commons licenses
> that are neither ShareAlike nor NoDerivs.
Whether that is "unfortunate" or not is, clearly, a matter of individual
perspective. I personally try to help people understand how licensing
works so they can implement their objectives. I have better things to
do than pass judgement on those objectives. If their idea of proper
licensing is too wacky for me, I will probably ignore their creations
and use something else.
> They don't state what kind of a license is required for the derivative
> work. I presume it means that one can change the license since only
> the ShareAlike ones prohibit changing the license.
ARGH! You cannot "change the licence" of something you don't own. You
can change the licence only of something you _do_ own -- by issuing a
second instance under different terms of your choosing.
Please don't talk about "changing the licence" (or "relicensing") of
things except when you're speaking of the copyright owner issuing new
instances. Those phrases have caused more confusion in licensing
discussions than any others -- even worse than referring to a licence as
a "copyright".
> But if one can change the license to anything, then one could turn a
> NonCommercial license into a Commercial one, etc which is a loophole
> in the Creative Commons license.
Let me try to turn that fuzzy and misleading paragraph into something
actually correct and clear:
Alice writes document foo, and issue an instance of it to the public
under the Noncommercial 1.0 licence. Bob downloads foo, and creates
document bar, which incorporates foo. Bob offers bar for sale, thinking
he's allowed to because bar is his property.
1. Has Bob infringed Alice's copyright? Clearly yes. Under clause 4b,
Bob...
"may not exercise any of the rights granted to [Bob] in
Section 3 above in any manner that is primarily intended for
or directed toward commercial advantage or private monetary
compensation."
2. Has Bob "changed the licensing" on Alice's work? No. Alice's work
is exactly as available as it was before, in exactly the same places.
The more interesting case is where Alice specifies Attribution 1.0
for foo, and Bob creates bar in obfuscated format that is hostile to
further derivatives, and forbids (under his copyright property interest
concerning his additions) redistribution by others so as to prop up
his price for selling bar under restrictive terms -- but still includes
an author credit for Alice.
1. Is Alice's copyright infringed? Clearly, no. Bob is satisfying
Alice's primary requirement under Attribution 1.0's clauses 4a and 4b,
with that author credit.
2. Is Bob denying the public access to Alice's work under her terms?
No. foo remains no more or less available than it always was, in the
exact same places as before. If it ever disappears from those places,
that's not Bob's fault -- or his responsibility.
3. Has Bob relicensed foo? No. He's specified differing (but
compatible) terms for his additional contribution of content, on an
instance of Alice's property.
4. Wait a minute! Isn't that relicensing of foo if the public can no
longer get to foo? No. The public was never guaranteed access to foo
in the first place. The public was granted the legal right to not be
sued for making basically any use of foo that includes Alice's author
credit. It wasn't guaranteed that everything incorporating foo would
always furnish foo in an optimally usable form.
In the same fashion, you're not guaranteed that anyone will give you
source code to a BSD-licensed codebase. If nobody's willing, then tough
luck for you: Nobody's rights have been violated. The licence states
only what you're allowed to do by law _if_ you can do it at all.
--
Cheers, "A raccoon tangled with a 23,000 volt line, today. The results
Rick Moen blacked out 1400 homes and, of course, one raccoon."
####@####.#### -- Steel City News
