Discussion:
Choosing a License: GNU APL? AFL 3.0?
Sean B. Palmer
2007-12-30 08:55:12 UTC
Permalink
Hello,

Noah Slater has offered to make a debian package from some of my
software, but I'm having trouble choosing a DFSG compatible license. I
was wondering if debian-legal could help.

I'm looking for a permissive license, of the Modified BSD or MIT
variety, but I'd like for the copyright notices in each file to be
protected without having to include the whole license itself, if it's
more than a few lines.

The Modified BSD, for example, only protects "the above copyright
notice, this list of conditions and the following disclaimer", so a
simple notice that says "Licensed under the Modified BSD License"
wouldn't be protected unless I included the whole license in each
file. Therefore it's not sufficient.

One of the best license suggestions I've had so far, from Noah
himself, is the GNU All Permissive License:

http://www.gnu.org/prep/maintain/html_node/License-Notices-for-Other-Files.html

But I'm not sure if this license is DFSG compatible, and it's not
listed as FSF or OSI approved (not that that's a surefire way of
telling whether something is DFSG compatible or not).

After a bit of poking about I decided to try my hand at making a
hybrid of the GNU APL and OSI approved Fair License; and in the
ensuing lengthy debate on OSI's license-review list, Larry Rosen
suggested that I use the Academic Free License 3.0.

I looked around a bit and found this said about the AFL 2.1:

"It seems that a rather clear consensus about the AFL is being formed
here at debian-legal: it doesn't comply with DFSG and it's
GPL-incompatible."
- http://lists.debian.org/debian-legal/2004/10/msg00263.html

Which is also my impression of how the AFL (2.1, at least) was
received here having read through the threads, but on the other hand
Rick Moen has told me that the AFL 3.0 is DFSG compliant:

"Both of those [OSL/AFL 3.0] comply with the DFSG."
- http://www.crynwr.com/cgi-bin/ezmlm-cgi?17:mmn:25

And he was willing to go into the nature of authority in the Debian
project--though I haven't yet taken up his offer of doing so, in
preference of actually, you know, asking debian-legal first :-)

To summarise my intent: I just want to add a couple of lines to each
file and be happy in the knowledge that people can do pretty much what
they want with my work except for removing the per-file attribution. I
like my packages to be modularly reusable and lean and pretty. The
license I choose has got to be unobtrusive and it's got to be a "keep
my name on this file" sort of thing; but beyond that, permissive.

I think that this is a FSF, OSI, and DFSG compatible requirement, but
I can't find a common license that supports it.

Perhaps the GNU APL is acceptable? Even its three lines is kinda long
for my tastes, however (I know, I know, I have ridiculous tastes). The
AFL 3.0 approach on the other hand seems to me a very good one, but
the license itself, in its 2.1 guise, is apparently very controversial
from the DFSG perspective. Also I was hoping for a more simple
BSD/MIT-like license text.

It's been suggested severally that I just use a notice to link to
whatever license I want, but remember that the linked license must in
that case protect the notice. Please read my badly explained
requirements carefully :-)

Probably I'm just hoping for the moon, but maybe someone can help me out here?

Thanks,
--
Sean B. Palmer, http://inamidst.com/sbp/
Francesco Poli
2007-12-30 10:32:02 UTC
Permalink
Post by Sean B. Palmer
Hello,
Hi!
Post by Sean B. Palmer
Noah Slater has offered to make a debian package from some of my
software, but I'm having trouble choosing a DFSG compatible license. I
was wondering if debian-legal could help.
I would be happy to help.
Post by Sean B. Palmer
I'm looking for a permissive license, of the Modified BSD or MIT
variety, but I'd like for the copyright notices in each file to be
protected without having to include the whole license itself, if it's
more than a few lines.
My usual recommendation for these needs is the Expat license:
http://www.jclark.com/xml/copying.txt
However, it seems that it is already too long for your tastes! :-O

[...]
Post by Sean B. Palmer
One of the best license suggestions I've had so far, from Noah
http://www.gnu.org/prep/maintain/html_node/License-Notices-for-Other-Files.html
I'll quote it here fully, for future reference:

| Copying and distribution of this file, with or without modification,
| are permitted in any medium without royalty provided the copyright
| notice and this notice are preserved.
Post by Sean B. Palmer
But I'm not sure if this license is DFSG compatible,
I don't see a reason why it should fail the DFSG.
I think a work solely licensed under the terms of this license does
fully comply with the DFSG.

Disclaimers: IANAL, TINLA, IANADD, TINASOTODP.
Post by Sean B. Palmer
After a bit of poking about I decided to try my hand at making a
hybrid of the GNU APL and OSI approved Fair License; and in the
ensuing lengthy debate on OSI's license-review list, Larry Rosen
suggested that I use the Academic Free License 3.0.
Writing new licenses is in general strongly recommended against,
because:

a) writing a good license is a long and very hard task
b) license proliferation is bad (since it adds to licensing complexity
and confusion, and balkanizes the Free Software community whenever
license incompatibilities occur) and should be avoided as often as
possible
Post by Sean B. Palmer
"It seems that a rather clear consensus about the AFL is being formed
here at debian-legal: it doesn't comply with DFSG and it's
GPL-incompatible."
- http://lists.debian.org/debian-legal/2004/10/msg00263.html
Frankly, I don't remember the details, but a quick glance of the cited
thread seems to reveal that I was not the only one that reached that
conclusion.
Post by Sean B. Palmer
Which is also my impression of how the AFL (2.1, at least) was
received here having read through the threads, but on the other hand
"Both of those [OSL/AFL 3.0] comply with the DFSG."
- http://www.crynwr.com/cgi-bin/ezmlm-cgi?17:mmn:25
And he was willing to go into the nature of authority in the Debian
project--though I haven't yet taken up his offer of doing so, in
preference of actually, you know, asking debian-legal first :-)
Well, in my understanding, debian-legal acts as a sort of advisory
board: we discuss about a topic, (hopefully) reach a consensus, and then
decision-makers (i.e.: FTP-masters) decide.

If Rick Moen means that, I think he's right, even though the politeness
of his message could have been improved...

I cannot comment on the AFL 3.0, since I haven't analyzed it.
Post by Sean B. Palmer
To summarise my intent: I just want to add a couple of lines to each
file and be happy in the knowledge that people can do pretty much what
they want with my work except for removing the per-file attribution. I
like my packages to be modularly reusable and lean and pretty. The
license I choose has got to be unobtrusive and it's got to be a "keep
my name on this file" sort of thing; but beyond that, permissive.
I think the simple license that you called "GNU All Permissive License"
satisfies your requirements.
If the 2-line-length constraint is so important, well, who said that
those two lines have to be 70 column long? You could use more columns
and make the license text fit in two lines! ;-)


[...]
Post by Sean B. Palmer
Probably I'm just hoping for the moon, but maybe someone can help me out here?
I hope this helps.
Post by Sean B. Palmer
Thanks,
You're welcome.
--
http://frx.netsons.org/doc/nanodocs/testing_workstation_install.html
Need to read a Debian testing installation walk-through?
..................................................... Francesco Poli .
GnuPG key fpr == C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4
Sean B. Palmer
2007-12-30 12:57:42 UTC
Permalink
Post by Francesco Poli
My usual recommendation for these needs is the Expat license
Which is indeed too long, even though James Clark is crazy-cool.
Post by Francesco Poli
Writing new licenses is in general strongly recommended against
Yeah, I understand. But it proved to be a pretty good way to start a
conversation, and I was interested to see what would happen. It was
also exceptionally fun, though perhaps I'm a rascal to say so.
Post by Francesco Poli
b) license proliferation is bad
Whilst that's very true, and I would prefer there to be far fewer
licenses than there are currently, my argument was that there simply
wasn't a license I could find that meets my (rather humble, I think!)
requirements. There's no harm in making new licenses if they meet new
and reasonable needs.

Wasn't GPLv3 conceived because of new demands with respect to software
patents and tivoisation and the like? Why were so few other licenses
revised in light of the same new demands, I wonder?

(This question is prolly getting out of scope; sorry.)
Post by Francesco Poli
I cannot comment on the AFL 3.0, since I haven't analyzed it.
Okay, no problem. It would be nice to see if there is a consensus,
either existing or newly emerging, about it though.
Post by Francesco Poli
I think the simple license that you called "GNU All Permissive
License" satisfies your requirements.
If the 2-line-length constraint is so important, well, who said that
those two lines have to be 70 column long?
Nobody; but Guido van Rossum says they must be 79 columns :-)

Really the difference comes down to the following:

* * *

Copyright 2007, Sean B. Palmer, inamidst.com
Released to the public under the Example License.

* * *

Compared to:

* * *

Copyright 2007, Sean B. Palmer, inamidst.com

Copying and distribution of this file, with or without modification, are
permitted in any medium without royalty provided the copyright notice and
this notice are preserved.

* * *

To most people these probably look identical, but the former is by far
the more preferable to me. It's my idiosyncrasy, and I'm gonna stick
by it! :-)

I'm also a bit concerned (my requirements have a habit of evolving)
about the warranty clause, or rather of the GNU APL's lack of one. I
could take the very short warranty disclaimer from the end of the OSI
approved Fair License, and tack it on as follows:

* * *

Copyright 2007, Sean B. Palmer, inamidst.com

Copying and distribution of this file, with or without modification, are
permitted in any medium without royalty provided the copyright notice and
this notice are preserved. Disclaimer: this work is without warranty.

* * *

But then we're possibly starting to get into novel license territory,
which as we've already seen is somewhat bad. Moreover the opinion of
whether you ought to have a warranty disclaimer seems to vary with who
you ask. I'm not sure how practically necessary it is.

Anyway, thanks for your opinion that the GNU APL is DFSG compatible!
--
Sean B. Palmer, http://inamidst.com/sbp/
Francesco Poli
2007-12-30 16:59:29 UTC
Permalink
[...]
Post by Sean B. Palmer
Post by Francesco Poli
b) license proliferation is bad
Whilst that's very true, and I would prefer there to be far fewer
licenses than there are currently, my argument was that there simply
wasn't a license I could find that meets my (rather humble, I think!)
requirements. There's no harm in making new licenses if they meet new
and reasonable needs.
Only when there are *really important* needs that are not even nearly
satisfied by existing licenses.
Please don't take offense for this, but I think that your needs are not
so critical that they cannot be bent a little to be satisfied by an
existing license.
Post by Sean B. Palmer
Wasn't GPLv3 conceived because of new demands with respect to software
patents and tivoisation and the like? Why were so few other licenses
revised in light of the same new demands, I wonder?
The GNU GPL v3 is worse than the GNU GPL v2 in my opinion: it has
several minor improvements over its predecessor, but implements a broken
copyleft in some respects.
Start here, if you are interested in the gory details:
http://lists.debian.org/debian-legal/2007/11/msg00104.html

[...]
Post by Sean B. Palmer
Post by Francesco Poli
I think the simple license that you called "GNU All Permissive
License" satisfies your requirements.
If the 2-line-length constraint is so important, well, who said that
those two lines have to be 70 column long?
Nobody; but Guido van Rossum says they must be 79 columns :-)
Is that a recommendation or a Python syntax rule? I don't remember any
such rule in Python and the following test script seems to work without
any complaints from the interpreter:


#!/usr/bin/python

# very long line, including copyright notice, permission notice, disclaimer of warranty, credits, acknowledgements, and some funny jokes...

print "Hello, world!"
Post by Sean B. Palmer
* * *
Copyright 2007, Sean B. Palmer, inamidst.com
Released to the public under the Example License.
* * *
* * *
Copyright 2007, Sean B. Palmer, inamidst.com
Copying and distribution of this file, with or without modification,
are permitted in any medium without royalty provided the copyright
notice and this notice are preserved.
* * *
To most people these probably look identical, but the former is by far
the more preferable to me. It's my idiosyncrasy, and I'm gonna stick
by it! :-)
Since you should place the Example License text somewhere accompanying
your source code (for clarity's sake, so that recipients know their
rights even in the absense of an Internet connection), I think the
decision boils down to: do you prefer

(A) having a short license included in all your source files

or rather

(B) having even shorter pointer to a separate file (e.g.: LICENSE.txt)
containing your more or less short license

?

[...]
Post by Sean B. Palmer
Anyway, thanks for your opinion that the GNU APL is DFSG compatible!
You're welcome.
Once again: please remember that this is my *personal* opinion.
IANAL, TINLA, IANADD, TINASOTODP.


P.S.:
Please do _not_ reply to my personal e-mail address, while Cc:ing the
list address, as I didn't ask you to do so.
Please follow the code of conduct on Debian lists:
http://www.debian.org/MailingLists/#codeofconduct
Thanks.
--
http://frx.netsons.org/doc/nanodocs/testing_workstation_install.html
Need to read a Debian testing installation walk-through?
..................................................... Francesco Poli .
GnuPG key fpr == C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4
Sean B. Palmer
2007-12-30 17:51:44 UTC
Permalink
Post by Francesco Poli
Please don't take offense for this, but I think that your needs are
not so critical that they cannot be bent a little to be satisfied by an
existing license.
Well several licenses have parts of what I need, but there's no
license which has all of what I need. I've found that the CDDL comes
closest on the OSI's "popular licenses" list:

A = Allows short statement of application
B = Preserves copyright statements and notices
C = Allows distribution without full license text
D = License is or may be fixed to exclude later versions

| A | B | C | D |
---------------------------+---+---+---+---+
Apache License 2.0 | N | Y | N | Y |
Modified BSD | Y | N | N | Y |
GNU GPL 2 | ? | ? | N | Y |
GNU LGPL 2.1 | ? | ? | N | N |
MIT | Y | N | N | Y |
Mozilla Public License 1.1 | N | Y | N | N |
CDDL 1.0 | Y | Y | N | Y |
Common Public License 1.0 | Y | Y | N | N |
---------------------------+---+---+---+---+

(This table may not be accurate. This is not legal advice.)

The only two problems with the CDDL from my perspective are that a) it
requires users to include a copy of it in all redistributions, and b)
though I wrote gave it a Y in the D column, it's actually not clear
how to make it only refer to a single version of the CDDL. I suspect
that the following might be okay:

* * *

Copyright 2007, Sean B. Palmer, inamidst.com
Released to the public under the CDDL, v1.0 only.

* * *

But that column C failure is a shame. Note that though every single
common license failed column C, some licenses such as the AFL 3.0 do
pass that column.

What I mean to show you with this table is that my requirements aren't
outlandish. It isn't that nobody has thought of them before, or that
nobody considers them important. It's that they're not covered all
together by any single license, as far as I can tell.
Post by Francesco Poli
Is that a recommendation or a Python syntax rule?
It's a recommendation that I always follow. If I didn't care about
code being lean and compact and pretty, then columns A and C wouldn't
be requirements for me! :-)
Post by Francesco Poli
Please do _not_ reply to my personal e-mail address
Whoops, sorry. I just hit "Reply to all" in my email client; "Reply"
would have sent it to you personally only. My email client is Google
Mail, so you might want to complain to its maintainer.

I will endeavour to strip personal email address from the recipients
whenever I send messages to debian-legal in future. Please forgive my
ignorance of this point of conduct, and any future instances where I
forget.

Thanks,
--
Sean B. Palmer, http://inamidst.com/sbp/
Sean B. Palmer
2007-12-30 18:22:57 UTC
Permalink
Post by Sean B. Palmer
A = Allows short statement of application
B = Preserves copyright statements and notices
C = Allows distribution without full license text
D = License is or may be fixed to exclude later versions
And here's the table augmented with what the OSI categorises as
"Other/Miscellaneous licenses" in its proliferation report:

| A | B | C | D |
----------------------------+---+---+---+---+
Apache License 2.0 | N | Y | N | Y |
Modified BSD | Y | N | N | Y |
GNU GPL 2 | ? | ? | N | Y |
GNU LGPL 2.1 | ? | ? | N | N |
MIT | Y | N | N | Y |
Mozilla Public License 1.1 | N | Y | N | N |
CDDL 1.0 | Y | Y | N | Y |
Common Public License 1.0 | Y | Y | N | N |
----------------------------+---+---+---+---+
Adaptive Public License 1.0 | N | Y | N | N |
The Artistic License | Y | Y | Y | Y |
Artistic License 2.0 | Y | N | Y | ? |
Open Software License 3.0 | Y | Y | Y | Y |
Q Public License 1.0 | Y | Y | N | Y |
zlib/libpng License | Y | N | ? | Y |
----------------------------+---+---+---+---+

(This table may not be accurate. This is not legal advice.)

The problem with The Artistic License is that, as Wikipedia says:

'Whether or not the original Artistic License is a free software
license is largely undecided. It was criticised by the Free Software
Foundation as being "too vague; some passages are too clever for their
own good, and their meaning is not clear."'
- http://en.wikipedia.org/wiki/Artistic_License

The problem with the OSL 3.0 is that it's the AFL 3.0's copyleft
sister, so DFSG problems with the AFL presumably affect the OSL 3.0
too, and it's not a permissive license. (It might be argued that no
license passing column B is a permissive license, but I think you
could make this argument for any license failing column C too, and yet
Modified BSD and MIT pass column C.)
--
Sean B. Palmer, http://inamidst.com/sbp/
MJ Ray
2007-12-31 01:47:38 UTC
Permalink
The only two problems with the CDDL from my perspective are [...]
I don't know whether the problems with the CDDL that 1. it does not
follow the DFSG if there are active patents on the Covered Software;
2. Matthew Garrett's comment at
http://lists.debian.org/debian-legal/2004/12/msg00004.html
about identification, (I'm not sure if 3.2 means the notice must identify
You against your wishes); are problems from your perspective. Also,
being a modified MPL, it imports any MPL problems and lawyerbombs.

[...]
Whoops, sorry. I just hit "Reply to all" in my email client; "Reply"
would have sent it to you personally only. My email client is Google
Mail, so you might want to complain to its maintainer.
Why don't you file a bug that Google Mail doesn't support List-Post and
other common mailing list behaviours? You're the user, after all.

Regards,
--
MJR/slef
My Opinion Only: see http://people.debian.org/~mjr/
Please follow http://www.uk.debian.org/MailingLists/#codeofconduct
Sean B. Palmer
2007-12-31 10:43:38 UTC
Permalink
Post by MJ Ray
I don't know whether the problems with the CDDL that 1. it does not
follow the DFSG if there are active patents on the Covered Software;
2. Matthew Garrett's comment at
http://lists.debian.org/debian-legal/2004/12/msg00004.html
Yes, thanks for bringing these to my attention. I'm currently thinking
about the Eiffel Forum License version 2, which is a pretty concise
license. It fails the column C test, but the single clause that makes
it fail:

* any distribution of this package, whether modified or not,
includes this license text.

- http://www.eiffel-nice.org/license/eiffel-forum-license-2.txt

Could easily be deleted. Is it DFSG compatible? Would it still be DFSG
compatible with this clause deleted?
Post by MJ Ray
Why don't you file a bug that Google Mail doesn't support List-Post and
other common mailing list behaviours? You're the user, after all.
Because I don't know that it's common, and I don't mind the behaviour
myself (I think that good clients should drop duplicates).

Thanks,
--
Sean B. Palmer, http://inamidst.com/sbp/
Francesco Poli
2008-01-01 10:18:50 UTC
Permalink
On Mon, 31 Dec 2007 10:43:38 +0000 Sean B. Palmer wrote:

[...]
Post by Sean B. Palmer
I'm currently thinking
about the Eiffel Forum License version 2, which is a pretty concise
license. It fails the column C test, but the single clause that makes
* any distribution of this package, whether modified or not,
includes this license text.
- http://www.eiffel-nice.org/license/eiffel-forum-license-2.txt
Could easily be deleted. Is it DFSG compatible? Would it still be DFSG
compatible with this clause deleted?
For future reference the full text of the Eiffel Forum License, version
2 is quoted below:


Eiffel Forum License, version 2

1. Permission is hereby granted to use, copy, modify and/or
distribute this package, provided that:
* copyright notices are retained unchanged,
* any distribution of this package, whether modified or not,
includes this license text.
2. Permission is hereby also granted to distribute binary programs
which depend on this package. If the binary program depends on a
modified version of this package, you are encouraged to publicly
release the modified version of this package.

***********************

THIS PACKAGE IS PROVIDED "AS IS" AND WITHOUT WARRANTY. ANY EXPRESS OR
IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE AUTHORS BE LIABLE TO ANY PARTY FOR ANY
DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES ARISING IN ANY WAY OUT OF THE USE OF THIS PACKAGE.

***********************


I think that works solely licensed under these terms comply with the
DFSG.
The reasons are: I don't see any clause in conflict with any DFSG; it's
compatible with the GNU GPL v2, as stated by the FSF
(http://www.gnu.org/licenses/license-list.html).

Moreover, there are (or have been) packages in Debian (main) licensed
under the Eiffel Forum License, version 2: one example is gobo
(http://packages.debian.org/source/etch/gobo).


Big disclaimers for this analysis: IANAL, TINLA, IANADD, TINASOTODP.
--
http://frx.netsons.org/progs/scripts/refresh-pubring.html
New! Version 0.6 available! What? See for yourself!
..................................................... Francesco Poli .
GnuPG key fpr == C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4
Sean B. Palmer
2008-01-01 15:29:38 UTC
Permalink
Post by Francesco Poli
I think that works solely licensed under these terms comply
with the DFSG.
Thanks. That concurs with other advice that I'd been given, and what
I'd been able to work out myself.

<aside>

I've had numerous comments, incidentally, that debian-legal is just a
useless discussion forum, a hive of horrors, a place to be avoided. I
know someone who's not subscribed to it on purpose. Someone else even
said outright that they're scared of it!

Perhaps I haven't been subscribed long enough or something, but I've
found the exact opposite. Certainly the feedback that I've got has
been tough and debatory. But it's also been prompt. It's been polite.
It's been terse and informative, and all emails I've received so far
have been, I think, written with the spirit of helpfulness.

So many thanks to everyone for that; and sorry to spoil your bad reputation :-)

</aside>
--
Sean B. Palmer, http://inamidst.com/sbp/
Ben Finney
2008-01-01 22:43:47 UTC
Permalink
Post by Sean B. Palmer
I've had numerous comments, incidentally, that debian-legal is just
a useless discussion forum, a hive of horrors, a place to be
avoided. I know someone who's not subscribed to it on purpose.
Someone else even said outright that they're scared of it!
Perhaps I haven't been subscribed long enough or something, but I've
found the exact opposite. Certainly the feedback that I've got has
been tough and debatory. But it's also been prompt. It's been
polite. It's been terse and informative, and all emails I've
received so far have been, I think, written with the spirit of
helpfulness.
This is very much a perception coloured by the participants (or
observers).

Those, like you, who choose to debate and allow their arguments and
assumptions to be examined in the interest of improving them, will
find debian-legal to be a more pleasant place.

Those who prefer *not* to have assumptions examined or authority
questioned will instead find debian-legal to be "a place to be
avoided".
Post by Sean B. Palmer
So many thanks to everyone for that; and sorry to spoil your bad reputation :-)
Happy to have some counterpoints made. Thanks!
--
\ "My aunt gave me a walkie-talkie for my birthday. She says if |
`\ I'm good, she'll give me the other one next year." -- Steven |
_o__) Wright |
Ben Finney
Marco d'Itri
2008-01-02 10:17:03 UTC
Permalink
Post by Sean B. Palmer
Perhaps I haven't been subscribed long enough or something, but I've
found the exact opposite. Certainly the feedback that I've got has
been tough and debatory. But it's also been prompt. It's been polite.
It's been terse and informative, and all emails I've received so far
have been, I think, written with the spirit of helpfulness.
You forgot the "being right" part, which some people believe to be quite
important...
--
ciao,
Marco
Francesco Poli
2008-01-02 17:10:40 UTC
Permalink
On Tue, 1 Jan 2008 15:29:38 +0000 Sean B. Palmer wrote:

[...]
Post by Sean B. Palmer
<aside>
I've had numerous comments, incidentally, that debian-legal is just a
useless discussion forum, a hive of horrors, a place to be avoided. I
know someone who's not subscribed to it on purpose. Someone else even
said outright that they're scared of it!
Perhaps I haven't been subscribed long enough or something, but I've
found the exact opposite. Certainly the feedback that I've got has
been tough and debatory. But it's also been prompt. It's been polite.
It's been terse and informative, and all emails I've received so far
have been, I think, written with the spirit of helpfulness.
So many thanks to everyone for that; and sorry to spoil your bad reputation :-)
</aside>
Well, thanks for this gratifying aside!
--
http://frx.netsons.org/progs/scripts/refresh-pubring.html
New! Version 0.6 available! What? See for yourself!
..................................................... Francesco Poli .
GnuPG key fpr == C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4
Francesco Poli
2007-12-31 11:36:52 UTC
Permalink
Post by Sean B. Palmer
Post by Francesco Poli
Please don't take offense for this, but I think that your needs are
not so critical that they cannot be bent a little to be satisfied by
an existing license.
Well several licenses have parts of what I need, but there's no
license which has all of what I need. I've found that the CDDL comes
The CDDL is a poor license choice: see
http://lists.debian.org/debian-legal/2006/12/msg00032.html
http://lists.debian.org/debian-legal/2005/09/msg00056.html

Remember: IANAL, TINLA, IANADD, TINASOTODP.
Post by Sean B. Palmer
A = Allows short statement of application
B = Preserves copyright statements and notices
C = Allows distribution without full license text
I am not able to see C as a reasonable requirement: including the full
license text when redistributing is something that should be done anyway
as a matter of clarity (so that recipients know their rights, even when
they don't have an Internet connection).
Hence, mandating it as a condition for the permission to reditribute
does not seem to be a bad thing.

Please reconsider your requirements: I would recommend you to drop C
entirely.
Post by Sean B. Palmer
D = License is or may be fixed to exclude later versions
D is a very agreeable requirement, from my point of view. I tend to
keep an eye on this feature, when I analyze licenses.
Post by Sean B. Palmer
| A | B | C | D |
---------------------------+---+---+---+---+
[...]
Post by Sean B. Palmer
Modified BSD | Y | N | N | Y |
[...]
Post by Sean B. Palmer
MIT | Y | N | N | Y |
Are we sure that B is N for the 3-clause BSD license and for the
Expat/MIT license?
If I understand requirement B correctly, I would say that B is Y for
these two licenses.

[...]
Post by Sean B. Palmer
But that column C failure is a shame.
No, it's a feature, I would say! ;-)

[...]
Post by Sean B. Palmer
Post by Francesco Poli
Please do _not_ reply to my personal e-mail address
Whoops, sorry. I just hit "Reply to all" in my email client; "Reply"
would have sent it to you personally only. My email client is Google
Mail, so you might want to complain to its maintainer.
No, I might want to complain to the users of that privacy bomb:
http://www.eff.org/deeplinks/archives/001398.php
http://www.epic.org/privacy/gmail/faq.html
http://www.gmail-is-too-creepy.com/
http://mjr.towers.org.uk/blog/2006/google

:-(
--
http://frx.netsons.org/doc/nanodocs/testing_workstation_install.html
Need to read a Debian testing installation walk-through?
..................................................... Francesco Poli .
GnuPG key fpr == C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4
Sean B. Palmer
2007-12-31 17:52:22 UTC
Permalink
Post by Francesco Poli
Post by Sean B. Palmer
C = Allows distribution without full license text
I am not able to see C as a reasonable requirement
Fair enough. I've made a note of that in the following article:

http://inamidst.com/stuff/eiffel/
- Why the Eiffel Forum License?

Which I've just been writing about why, if you don't bother taking
column C into account, the rather obscure but awesome Eiffel Forum
License comes out on top.

Note that the EFL v.2 is OSI and FSF approved, and GPL compatible. And
best of all, for my direct needs, it's DFSG compliant and meets
everything apart from column C. Seems like a winner, unless anyone can
suggest why not!
Post by Francesco Poli
Are we sure that B is N for the 3-clause BSD license and for
the Expat/MIT license?
Yes, there is no text in those licenses which requires that all
copyright notices within the package they apply to are retained and
preserved.
Post by Francesco Poli
http://www.eff.org/deeplinks/archives/001398.php
http://www.epic.org/privacy/gmail/faq.html
http://www.gmail-is-too-creepy.com/
http://mjr.towers.org.uk/blog/2006/google
Well at least you're not forced to use Google Mail.

Not yet, anyway.

(Muahahaha!)
--
Sean B. Palmer, http://inamidst.com/sbp/
Francesco Poli
2008-01-01 10:39:01 UTC
Permalink
Post by Sean B. Palmer
Post by Francesco Poli
Post by Sean B. Palmer
C = Allows distribution without full license text
I am not able to see C as a reasonable requirement
http://inamidst.com/stuff/eiffel/
- Why the Eiffel Forum License?
I've just read it (typo: my name is Francesco, not Francisco...).
I'm not convinced about the preserving copyright notices issue, more on
this later.

[...]
Post by Sean B. Palmer
Post by Francesco Poli
Are we sure that B is N for the 3-clause BSD license and for
the Expat/MIT license?
Yes, there is no text in those licenses which requires that all
copyright notices within the package they apply to are retained and
preserved.
IIUC, your concern is with the word "above".
The Expat license[1] states, in part:

| The above copyright notice and this permission notice shall be
| included in all copies or substantial portions of the Software.

[1] http://www.jclark.com/xml/copying.txt

The 3-clause BSD license[2] states, in part:

| * Redistributions of source code must retain the above copyright
| notice, this list of conditions and the following disclaimer.

[2] http://www.gnu.org/licenses/info/BSD_3Clause.html

You think that these clauses only apply to copyright notices that are
placed *above* the license text, while giving permission to strip or
alter isolated copyright notices that are near a line that just refers
to the license text. Did I get what you mean?

I'm not convinced that this is the case.
Firstoff, does the law allow removing or altering appropriate copyright
notices? I don't think so, but I'll leave this argument to real
lawyers...
Secondly, I think that the license text says "the above copyright
notice" just because it is designed to be included verbatim in each
file. However, if you prefer placing the license text in a centralized
file (e.g.: LICENSE.txt) and just put copyright notices (+ a line
pointing to the centralized file) in each source file, that's just a
matter of practical convenience: I don't think that this could alter the
meaning of the license, which mandates the preservation of copyright
notices...


To conclude, I still recommend you to adopt the Expat license (even
though the Eiffel Forum License, version 2 is acceptable).

Disclaimers: IANAL, TINLA, IANADD, TINASOTODP.
--
http://frx.netsons.org/progs/scripts/refresh-pubring.html
New! Version 0.6 available! What? See for yourself!
..................................................... Francesco Poli .
GnuPG key fpr == C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4
Sean B. Palmer
2008-01-01 15:20:27 UTC
Permalink
Post by Francesco Poli
(typo: my name is Francesco, not Francisco...)
Argh, sorry. Too much Hamlet!

I've expanded the article this morning, and corrected the typo.
Post by Francesco Poli
You think that these clauses only apply to copyright
notices that are placed *above* the license text, while
giving permission to strip or alter isolated copyright
notices that are near a line that just refers to the
license text. Did I get what you mean?
Yep.
Post by Francesco Poli
I'm not convinced that this is the case.
Firstoff, does the law allow removing or altering appropriate
copyright notices? I don't think so, but I'll leave this argument
to real lawyers...
The argument against this is that some licenses, all those that have a
Y in Column B of my survey, require this explicitly and some don't.
None, however, qualify it in terms of a universal restriction that the
licenses are simply reminding you of.

This is the heuristic argument I used with myself in lieu of being a
lawyer. In fact, I didn't think about it much—it just seemed like
common sense to me.

At the very least, a license that makes this point explicit makes me
feel better that a requirement I care about is going to be noticed by
people using my software.
Post by Francesco Poli
Secondly, I think that the license text says "the above copyright
notice" just because it is designed to be included verbatim in
each file.
People don't include it verbatim in each file, though. James Clark
doesn't do that in Expat, for example!

"See the file copying.txt for copying permission."
- e.g. expat/xmlparse/xmlparse.c

The requirement text in the Expat license is:

"The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software."

The *nature* of required inclusion isn't elucidated upon.
Post by Francesco Poli
LICENSE.txt) and just put copyright notices (+ a line pointing
to the centralized file) in each source file, that's just a matter
of practical convenience
But then the license only requires that the copyright notice and the
permission notice be preserved, which means the LICENSE.txt file.
Post by Francesco Poli
I don't think that this could alter the meaning of the license, which
mandates the preservation of copyright notices...
No, it says the "above copyright notice and this permission notice"
only. When you're using a LICENSE.txt file, that means the content of
the LICENSE.txt file only. It doesn't say anything about anything
else.
Post by Francesco Poli
Disclaimers: IANAL, TINLA, IANADD, TINASOTODP.
Same here! :-)
--
Sean B. Palmer, http://inamidst.co
Ben Finney
2008-01-01 22:36:26 UTC
Permalink
Post by Sean B. Palmer
Post by Francesco Poli
Firstoff, does the law allow removing or altering appropriate
copyright notices? I don't think so, but I'll leave this argument
to real lawyers...
The argument against this is that some licenses, all those that have a
Y in Column B of my survey, require this explicitly and some don't.
That doesn't seem to be an argument against Francesco's point at all.
A copyright license doesn't need to make anything explicit that is
already explicit in copyright law.

Some license authors may choose to make some requirements explicit
anyway, as a courtesy; but there is a huge body of requirement that
they omit from such licenses, that nevertheless continues to have
force of law.
Post by Sean B. Palmer
At the very least, a license that makes this point explicit makes me
feel better that a requirement I care about is going to be noticed
by people using my software.
I wonder how many other requirements that you care about in copyright
law are entirely unexamined because you've not seen them explicated in
any license text.
--
\ "Not using Microsoft products is like being a non-smoker 40 or |
`\ 50 years ago: You can choose not to smoke, yourself, but it's |
_o__) hard to avoid second-hand smoke." -- Michael Tiemann |
Ben Finney
Sean B. Palmer
2008-01-02 11:12:08 UTC
Permalink
Post by Ben Finney
A copyright license doesn't need to make anything explicit that
is already explicit in copyright law.
Okay. So is the preservation of copyright notices already explicit in
copyright law? I haven't been able to find anything about this on
Google and Wikipedia etc., though it's hard to know what to look for
with the question being the fairly subtle one of whether unqualified
permission to modify a file *entails* modification of copyright
notices.
Post by Ben Finney
Some license authors may choose to make some requirements
explicit anyway, as a courtesy
That's true, but in such cases I normally find qualifying text such as
"your right to ?thing is unaffected". On the other hand it's also true
that this is more often the case for making rights explicit than
making responsibilities explicit.
Post by Ben Finney
I wonder how many other requirements that you care about in
copyright law are entirely unexamined because you've not seen
them explicated in any license text.
Well, what other heuristic should I use? I can't justify the cost of a
lawyer for this, so I can only make the best assumptions that I know
how to make, and ask friends and people on debian-legal and so on...

In other words, if you have evidence that's better than mine, please
do tell me!
--
Sean B. Palmer, http://inamidst.com/sbp/
John Halton
2008-01-02 11:23:21 UTC
Permalink
Post by Sean B. Palmer
Okay. So is the preservation of copyright notices already explicit in
copyright law?
I'm pretty sure the answer to that would be "no", not least because
most jurisdictions have no requirements for copyright notices anyway
(copyright arising automatically without any formalities).

So if someone gives you blanket permission to modify a work, then that
would include the right to modify the copyright notices. However,
actually doing so may then conflict with other rights the author may
have, separate from copyright - e.g. "moral rights" giving an author
to be identified as such.

Plus, a general legal requirement to preserve copyright notices would
also require a general legal definition of "copyright notice", which
would not necessarily be a simple matter. If I put the following in a
copyright work: "(C) 2007 John Halton. All rights reserved. Licensed
under GPL v.3 or later. Absolutely no warranty applies to the maximum
attempt possible by law. Copies of any modifications would be
appreciated but are not required. Bespoke development work may be
carried out by author on request - please contact me for details.
Debian rocks!!!", at what point has that ceased to be a "copyright
notice" and become something else?

John

(TINLA)
Sean B. Palmer
2008-01-02 11:42:59 UTC
Permalink
Post by Sean B. Palmer
the question being the fairly subtle one of whether unqualified
permission to modify a file *entails* modification of copyright
notices.
Ah, this may in fact be irrelevant!

"The phrase /above copyright notice/ is somewhat misleading.
Presumably the BSD license really refers to the actual copy-
right notice that is displayed on the software being distributed
rather than the copyright notice shown above in the license,
for otherwise this would be a meaningless requirement."

- Larry Rosen, http://www.rosenlaw.com/Rosen_Ch05.pdf (p.80)

Presumably this applies to the Expat/MIT license too.
--
Sean B. Palmer, http://inamidst.com/sbp/
Ben Finney
2008-01-02 13:53:18 UTC
Permalink
Post by Sean B. Palmer
Post by Ben Finney
A copyright license doesn't need to make anything explicit that
is already explicit in copyright law.
Okay. So is the preservation of copyright notices already explicit
in copyright law?
I don't know. My current understanding is "yes" to a first
approximation. It would take actual lawyers to provide a more
trustworthy answer on this one; and it's very likely the answer
involves "it depends".
Post by Sean B. Palmer
Post by Ben Finney
I wonder how many other requirements that you care about in
copyright law are entirely unexamined because you've not seen them
explicated in any license text.
Well, what other heuristic should I use? I can't justify the cost of
a lawyer for this, so I can only make the best assumptions that I
know how to make, and ask friends and people on debian-legal and so
on...
It is to your credit that you do so, seriously.
--
\ "I know you believe you understood what you think I said, but I |
`\ am not sure you realize that what you heard is not what I |
_o__) meant." -- Robert J. McCloskey |
Ben Finney
Arnoud Engelfriet
2008-01-02 15:05:31 UTC
Permalink
Post by Ben Finney
Post by Sean B. Palmer
Okay. So is the preservation of copyright notices already explicit
in copyright law?
I don't know. My current understanding is "yes" to a first
approximation. It would take actual lawyers to provide a more
trustworthy answer on this one; and it's very likely the answer
involves "it depends".
Ok, you're a lawyer now. :)

Art. 6bis of Berne provides the right of the author to claim
authorship of the work. It seems reasonable to consider a copyright
notice as a claim of authorship (in the normal case, where
author == copyright holder). This "claim" includes the right to
be mentioned as the author. If the author has exercised this
right, removing the notice from the work violates this right.

Under US law, art. 6bis Berne does not apply (even though it should)
to authors of computer programs.

Arnoud
--
Arnoud Engelfriet, Dutch & European patent attorney - Speaking only for myself
Patents, copyright and IPR explained for techies: http://www.iusmentis.com/
Arnoud blogt nu ook: http://blog.iusmentis.com/
John Halton
2008-01-02 15:20:35 UTC
Permalink
Post by Arnoud Engelfriet
Art. 6bis of Berne provides the right of the author to claim
authorship of the work. It seems reasonable to consider a copyright
notice as a claim of authorship (in the normal case, where
author == copyright holder). This "claim" includes the right to
be mentioned as the author. If the author has exercised this
right, removing the notice from the work violates this right.
A couple of thoughts on this:

a. I don't think a copyright notice can be taken as an assertion of
authorship, precisely because the copyright may be owned by someone
other than the author. Hence most books that are published in the UK
include both a copyright notice (usually naming the publisher as
owner) and a statement that the author has asserted their moral
rights.

b. Even if a copyright notice could be read as an assertion of the
author's moral rights to be identified as the author, I don't think
this would make the text of the notice itself inviolable - provided
that the author was still acknowledged elsewhere.
Post by Arnoud Engelfriet
Under US law, art. 6bis Berne does not apply (even though it should)
to authors of computer programs.
Ditto English law. I think that betrays an assumption on the part of
US and UK lawmakers that computer programmers are merely technicians,
not creative artists.

John

(TINLA)
Arnoud Engelfriet
2008-01-02 15:38:28 UTC
Permalink
Post by John Halton
a. I don't think a copyright notice can be taken as an assertion of
authorship, precisely because the copyright may be owned by someone
other than the author. Hence most books that are published in the UK
include both a copyright notice (usually naming the publisher as
owner) and a statement that the author has asserted their moral
rights.
True. Let's try again: if the sole mention of a natural person who
is the creator occurs in the form of a copyright notice, then I
would take that as a claim of authorship.
Post by John Halton
b. Even if a copyright notice could be read as an assertion of the
author's moral rights to be identified as the author, I don't think
this would make the text of the notice itself inviolable - provided
that the author was still acknowledged elsewhere.
I'm not sure. If the author chooses a particular way of being
mentioned as author, does that override any other way of identifying
him as author?
Post by John Halton
Post by Arnoud Engelfriet
Under US law, art. 6bis Berne does not apply (even though it should)
to authors of computer programs.
Ditto English law. I think that betrays an assumption on the part of
US and UK lawmakers that computer programmers are merely technicians,
not creative artists.
Can Berne be invoked directly in the UK? I know the US has closed
that route. The issue here is that computer programs are to be
protected as "literary works" in Europe (Directive 91/250/EC), and
that means they get the full range of protection. Moral rights are
part of that protection, so a EU member state should have the
moral rights available to software authors.

Arnoud
--
Arnoud Engelfriet, Dutch & European patent attorney - Speaking only for myself
Patents, copyright and IPR explained for techies: http://www.iusmentis.com/
Arnoud blogt nu ook: http://blog.iusmentis.com/
Francesco Poli
2008-01-02 17:32:07 UTC
Permalink
Post by Sean B. Palmer
Post by Francesco Poli
(typo: my name is Francesco, not Francisco...)
Argh, sorry. Too much Hamlet!
Heh! ;-)
Post by Sean B. Palmer
I've expanded the article this morning, and corrected the typo.
I haven't had the time to fully re-read it, but I noticed something: the
document is released under the terms of CC-by-nd-v2.0, which is utterly
non-free.
I wonder why you want to be so permissive with your programs, while you
decide to be so restrictive with this document... :-(

I always recommend against any Creative Commons license.
v1.0, v2.0 and v2.5 licenses fail to meet the DFSG:
http://lists.debian.org/debian-legal/2004/04/msg00031.html
http://people.debian.org/~evan/ccsummary.html
My opinion is that v3.0 licenses also fail to meet the DFSG:
http://lists.debian.org/debian-legal/2007/03/msg00105.html
http://lists.debian.org/debian-legal/2007/09/msg00176.html
On the other hand, other people (including the FTP masters) seem to
disagree with me and think that CC-by-v3.0 and CC-by-sa-v3.0 meet the
DFSG... :-(


Disclaimers again: IANAL, TINLA, IANADD, TINASOTODP.


[...]
Post by Sean B. Palmer
Post by Francesco Poli
Secondly, I think that the license text says "the above copyright
notice" just because it is designed to be included verbatim in
each file.
People don't include it verbatim in each file, though. James Clark
doesn't do that in Expat, for example!
"See the file copying.txt for copying permission."
- e.g. expat/xmlparse/xmlparse.c
Interesting, I wasn't aware of that...
--
http://frx.netsons.org/progs/scripts/refresh-pubring.html
New! Version 0.6 available! What? See for yourself!
..................................................... Francesco Poli .
GnuPG key fpr == C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4
MJ Ray
2008-01-02 01:12:37 UTC
Permalink
Post by Sean B. Palmer
Well at least you're not forced to use Google Mail.
We seem to be forced to deal with their unrepentant users who make
it send base64'ed posts to this list with very long and spammy
headers, ignoring List-Post and the list code of conduct, yet
refusing to file bug reports with their software supplier.

What's the easiest way? k for killfile, sorry. I'm out.

Have they fixed that google bug where you can hide whole threads by
sending a known spam body with forged list References and similar
headers to the googlemail-using subscriber?

Regards,
--
MJR/slef
My Opinion Only: see http://people.debian.org/~mjr/
Please follow http://www.uk.debian.org/MailingLists/#codeofconduct
Ben Finney
2007-12-31 01:46:08 UTC
Permalink
Post by Sean B. Palmer
Post by Francesco Poli
Writing new licenses is in general strongly recommended against
Yeah, I understand. But it proved to be a pretty good way to start a
conversation, and I was interested to see what would happen. It was
also exceptionally fun, though perhaps I'm a rascal to say so.
You're not a rascal to say so: it *is* fun creating new things, and
the law can sometimes be a fun toolkit with which to play with
building things.

The problem is that, while proliferation of diverse works is generally
good, proliferation of diverse license terms for those works is mostly
bad for everyone involved. Bringing a newly-cobbled set of license
terms here for discussion usually implies those license terms are or
will be applied to a work, which is generally a bad thing.
--
\ "As the evening sky faded from a salmon color to a sort of |
`\ flint gray, I thought back to the salmon I caught that morning, |
_o__) and how gray he was, and how I named him Flint." -- Jack Handey |
Ben Finney
Ben Finney
2007-12-30 10:58:15 UTC
Permalink
Post by Sean B. Palmer
Noah Slater has offered to make a debian package from some of my
software, but I'm having trouble choosing a DFSG compatible license.
Thanks for your desire to contribute, and for your effort in gathering
information before choosing license terms for your work.
Post by Sean B. Palmer
I'm looking for a permissive license, of the Modified BSD or MIT
variety, but I'd like for the copyright notices in each file to be
protected without having to include the whole license itself, if
it's more than a few lines.
I'm not sure I take your meaning in "copyright notices" being
"protected". Copyright law protects any creative work (with
jurisdiction-specific exceptions), not the legal notice on that work.
Post by Sean B. Palmer
The Modified BSD, for example, only protects "the above copyright
notice, this list of conditions and the following disclaimer"
No, the 3-clause BSD license (like any copyright license) *grants
permission* to perform acts that are otherwise restricted by copyright
law. It is copyright law that, in the absence of any license, protects
the entire work.

The clause you're referring to is part of a condition on exercise of
the permissions granted by the license: that the specific texts it
refers to must be retained in the redistributed work.

I'm not sure how you got the idea that only certain parts of the work
are restricted; it's copyright law that does the restricting, *not*
the license terms.
Post by Sean B. Palmer
so a simple notice that says "Licensed under the Modified BSD
License" wouldn't be protected unless I included the whole license
in each file. Therefore it's not sufficient.
I'm fairly sure you'll find many examples of works that do exactly
that. So long as the license terms are explicit for every part of the
work, and unambiguous once received, the Debian project treats them as
the actual license terms for the work.

It's best, of course, to ensure the license terms themselves are
distributed along with the work; but this doesn't mean they need to be
part of every single file in the work. The instructions for use
accompanying the GNU GPL are the classic example of how to indicate
the license terms in every file without *including* the license terms
in every file.
Post by Sean B. Palmer
After a bit of poking about I decided to try my hand at making a
hybrid of the GNU APL and OSI approved Fair License
This is usually ill-advised. Please check again whether the license
terms of any of the popular, well-understood free-software licenses
are acceptable. The 3-clause BSD license that you already like would
be fine, and can be *applied* to every file without necessarily being
*included* in every file. Merely *referencing* the exact license terms
unambiguously in each file should be sufficient.
Post by Sean B. Palmer
It's been suggested severally that I just use a notice to link to
whatever license I want, but remember that the linked license must
in that case protect the notice. Please read my badly explained
requirements carefully :-)
I hope I've explained above that it's not the *license* that does
this, but copyright law itself. You, as copyright holder in your work,
are free to choose license terms and put an appropriate copyright
notice in your files when you distribute them. If you grant permission
to redistribute at all, it's copyright law that requires that notice
to be included in any redistributions of that work until you
explicitly give permission to the contrary.
--
\ "[...] a Microsoft Certified System Engineer is to information |
`\ technology as a McDonalds Certified Food Specialist is to the |
_o__) culinary arts." —Michael Bacarella |
Ben Finney
--
To UNSUBSCRIBE, email to debian-legal-***@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact ***@lists.debian.org
Arnoud Engelfriet
2007-12-30 11:42:02 UTC
Permalink
Post by Ben Finney
I hope I've explained above that it's not the *license* that does
this, but copyright law itself. You, as copyright holder in your work,
are free to choose license terms and put an appropriate copyright
notice in your files when you distribute them. If you grant permission
to redistribute at all, it's copyright law that requires that notice
to be included in any redistributions of that work until you
explicitly give permission to the contrary.
One case where this could become problematic is when permission is
granted to create derivative works. If the derivative work can be
distributed in binary-only form, then the copyright notices in the
source code become irrelevant. In such a case, a requirement like
article 2 of the BSD license to put a notice in the documentation
would be a good idea.

Arnoud
--
Arnoud Engelfriet, Dutch & European patent attorney - Speaking only for myself
Patents, copyright and IPR explained for techies: http://www.iusmentis.com/
Arnoud blogt nu ook: http://blog.iusmentis.com/
Ben Finney
2007-12-30 12:49:06 UTC
Permalink
Post by Arnoud Engelfriet
If you grant permission to redistribute at all, it's copyright law
that requires [the copyright notice] to be included in any
redistributions of that work until you explicitly give permission
to the contrary.
One case where this could become problematic is when permission is
granted to create derivative works. If the derivative work can be
distributed in binary-only form, then the copyright notices in the
source code become irrelevant.
I'm not sure how copyright law treats this in general. Is it the case
that a binary form, compiled from a copyrighted source form of a work,
is a "derived work", or is it the original copyrighted work itself?

I imagine that, whatever the answer to the above question -- whether
the redistribution in binary form is either the original work or a
derived work of the original -- the original copyright notice still
*applies*, whether it's included or not.
Post by Arnoud Engelfriet
In such a case, a requirement like article 2 of the BSD license to
put a notice in the documentation would be a good idea.
I prefer the GPLv2's section 1, which doesn't distinguish between
different forms of the work. It simply requires:

... that you conspicuously and appropriately publish on each copy
an appropriate copyright notice and disclaimer of warranty; keep
intact all the notices that refer to this License and to the
absence of any warranty; and give any other recipients of the
Program a copy of this License along with the Program.

This requires the notices to continue to apply to and accompany *any*
form of the work ("the Program", in GPL definitions) when
redistributed.
--
\ "If [a technology company] has confidence in their future |
`\ ability to innovate, the importance they place on protecting |
_o__) their past innovations really should decline." —Gary Barnett |
Ben Finney
--
To UNSUBSCRIBE, email to debian-legal-***@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact ***@lists.debian.org
Arnoud Engelfriet
2007-12-30 13:12:00 UTC
Permalink
Post by Ben Finney
Post by Arnoud Engelfriet
One case where this could become problematic is when permission is
granted to create derivative works. If the derivative work can be
distributed in binary-only form, then the copyright notices in the
source code become irrelevant.
I'm not sure how copyright law treats this in general. Is it the case
that a binary form, compiled from a copyrighted source form of a work,
is a "derived work", or is it the original copyrighted work itself?
I'd argue it is a translation and therefore a derivative work.

The case becomes a little more clear when the source code is edited
and/or additional pieces of software are linked into the binary.
Post by Ben Finney
I imagine that, whatever the answer to the above question -- whether
the redistribution in binary form is either the original work or a
derived work of the original -- the original copyright notice still
*applies*, whether it's included or not.
Of course. Copyright remains in force until 70 years after the
author has died (etc) regardless of any formalities.

Arnoud
--
Arnoud Engelfriet, Dutch & European patent attorney - Speaking only for myself
Patents, copyright and IPR explained for techies: http://www.iusmentis.com/
Arnoud blogt nu ook: http://blog.iusmentis.com/
Francesco Poli
2007-12-30 16:26:49 UTC
Permalink
Post by Arnoud Engelfriet
Post by Ben Finney
Post by Arnoud Engelfriet
One case where this could become problematic is when permission is
granted to create derivative works. If the derivative work can be
distributed in binary-only form, then the copyright notices in the
source code become irrelevant.
I'm not sure how copyright law treats this in general. Is it the
case that a binary form, compiled from a copyrighted source form of
a work, is a "derived work", or is it the original copyrighted work
itself?
I'd argue it is a translation and therefore a derivative work.
I was under the impression that a "mechanical" (i.e.: automated, without
any new creative input from the compiler user) translation didn't create
a derivative work, just a different form of the same work.
Compare with photocopying a piece of paper.

By contrast, when I manually translate a work (from a programming or
natural language into another one), I create a derivative.

Of course, the compiler may introduce some copyrighted material (runtime
support and the like) into the binary. In that case, the binary would
be a derivative of the source code and of the added material...


Was I under a wrong impression?
Post by Arnoud Engelfriet
The case becomes a little more clear when the source code is edited
This is clearcut: I would say that a derivative is formed, as long as
the editing adds enough creativity.
Is that right?
Post by Arnoud Engelfriet
and/or additional pieces of software are linked into the binary.
Let's not get started on the FSF's legal theory of linking: this belongs
in a separate thread, if you are willing to discuss its merits.


My usual disclaimers: IANAL, TINLA, IANADD, TINASOTODP.
--
http://frx.netsons.org/doc/nanodocs/testing_workstation_install.html
Need to read a Debian testing installation walk-through?
..................................................... Francesco Poli .
GnuPG key fpr == C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4
Arnoud Engelfriet
2007-12-30 18:35:51 UTC
Permalink
Post by Francesco Poli
Post by Arnoud Engelfriet
I'd argue it is a translation and therefore a derivative work.
I was under the impression that a "mechanical" (i.e.: automated, without
any new creative input from the compiler user) translation didn't create
a derivative work, just a different form of the same work.
Compare with photocopying a piece of paper.
Good point. Under US law you're probably right. Other countries sometimes
group anything that's not a literal reproduction under 'derivative work'.

In those jurisdictions (mine -the Netherlands- for one) you then have
non-creative and creative derivative works.

If I compile source into binary, I have a non-creative derivative.
If I edit the source and compile it, I have a creative derivative.
The difference is that I have a copyright on the latter but not
on the former, thanks to my creative additions to the latter.

So the question is how "derivative" is defined under local law.

Arnoud
--
Arnoud Engelfriet, Dutch & European patent attorney - Speaking only for myself
Patents, copyright and IPR explained for techies: http://www.iusmentis.com/
Arnoud blogt nu ook: http://blog.iusmentis.com/
Sean B. Palmer
2007-12-30 13:19:29 UTC
Permalink
Post by Ben Finney
I'm not sure I take your meaning in "copyright notices" being
"protected". Copyright law protects any creative work (with
jurisdiction-specific exceptions), not the legal notice on that work.
The AFL 3.0, for example, states:

'This Academic Free License (the "License") applies to any original
work of authorship (the "Original Work") whose owner (the "Licensor")
has placed the following licensing notice adjacent to the copyright
notice for the Original Work:

Licensed under the Academic Free License version 3.0'
- http://opensource.org/licenses/academic.php

So I could use that as follows:

* * *

#!/bin/sh
# Copyright 2007, Sean B. Palmer, inamidst.com
# Licensed under the Academic Free License version 3.0.

[&c.]

* * *

And now you can make modifications to the file and redistribute it,
and you don't have to include a copy of the Academic Free License
anywhere. But you can never strip the copyright line or the notice
line from the file.

Now say that I used Modified BSD instead.

* * *

#!/bin/sh
# Copyright 2007, Sean B. Palmer, inamidst.com
# Licensed under the Modified BSD License.

[&c.]

* * *

There are now two differences:

* Whenever you redistribute the file, you have to include a copy of
the Modified BSD License, either in the file itself or alongside it.
* You can technically strip the copyright line and the notice line
from the file and move them to a LICENSE.txt file instead. Not that
you would, but you could.

This is kinda crummy for you because you have to include a copy of the
Modified BSD license now, and it's kinda crummy for any subsequent
users and me the original author if you move the notice out because
then nobody knows who wrote it unless they look at the LICENSE.txt
file.

I think of licenses as mandating politeness and good academic
behaviour. They also let you know what the original author thinks of
as being polite :-)

So this is what I mean by preserving the notice. I hope that explains
it a bit better, but please feel free to point out anything I've
misunderstood or to ask further questions of course. Thanks for your
reply.
--
Sean B. Palmer, http://inamidst.com/sbp/
Ben Finney
2007-12-31 02:00:21 UTC
Permalink
Sean, please follow the Debian mailing list guidelines
<URL:http://www.debian.org/MailingLists/#codeofconduct>; in
particular, please don't send personal copies of messages also sent to
the list unless they're explicitly requested.
Post by Ben Finney
I'm not sure I take your meaning in "copyright notices" being
"protected". Copyright law protects any creative work (with
jurisdiction-specific exceptions), not the legal notice on that work.
The AFL 3.0, for example [...]
And now you can make modifications to the file and redistribute it,
and you don't have to include a copy of the Academic Free License
anywhere. But you can never strip the copyright line or the notice
line from the file.
I'm not convinced that's true. Copyright law, in my understanding,
requires that the recipient must have explicit license terms, or they
have no license by default.

It's not necessary to *include* the license terms when redistributing,
though that's surely a no-brainer way to make them explicit.
Now say that I used Modified BSD instead. [...]
* Whenever you redistribute the file, you have to include a copy of
the Modified BSD License, either in the file itself or alongside it.
This would appear to be *effectively* true for any license (whether
for a free or non-free work) that allows redistribution. Even if the
license terms don't state the requirement, my understanding is that
the effect of copyright law makes it mandatory to make license terms
explicit to the recipient of such a work in order to have any license
in the work.
I think of licenses as mandating politeness and good academic
behaviour. They also let you know what the original author thinks of
as being polite :-)
This may not be a good analogy to work from. There are a great many
license terms that are anything but "politeness and academic
behaviour"; I know many people who agree openly that the license terms
under which they distribute their software is greatly divergent from
their norm of politeness. (They resolve that disconnect by introducing
non-sequiturs about being compensated for their work, but that's not a
discussion that belongs in this thread.)

Instead, a better model would be to think of licenses as grants of
largesse from a monopoly holder, under whatever terms they choose.
--
\ "I think a good gift for the President would be a chocolate |
`\ revolver. And since he's so busy, you'd probably have to run up |
_o__) to him real quick and hand it to him." -- Jack Handey |
Ben Finney
Sean B. Palmer
2007-12-31 10:51:41 UTC
Permalink
Post by Ben Finney
Sean, please follow the Debian mailing list guidelines
<URL:http://www.debian.org/MailingLists/#codeofconduct>
Thanks. Francisco Poli pointed this out too, and I replied to him in
the last two paragraphs of this message:

http://lists.debian.org/debian-legal/2007/12/msg00115.html
Post by Ben Finney
Copyright law, in my understanding, requires that the
recipient must have explicit license terms, or they have
no license by default.
Yes, but I've been told that the expression of explicit license terms
can have a very compact expression:

http://lists.w3.org/Archives/Public/www-archive/2007Dec/0110

Here, Larry Rosen argues that you can simply refer to any OSI or CC
approved license (he didn't mention the FSF explicitly; I wonder if
FSF approved licenses can be referred to as well? :-) by name, in a
one line statement below the copyright notice.
Post by Ben Finney
Instead, a better model would be to think of licenses as
grants of largesse from a monopoly holder, under whatever
terms they choose.
Heh, thanks, I like that.

The Eiffel Forum License v2 appears to be very close to my
requirements, as I recently noted in a previous email in this thread.
I would need to delete a single clause to, I think, make it fully
compatible with my requirements. Is it DFSG compatible in your
opinion? It doesn't seem to be widely used, sadly.
--
Sean B. Palmer, http://inamidst.com/sbp/
Josselin Mouette
2008-01-02 16:04:28 UTC
Permalink
Post by Sean B. Palmer
I'm looking for a permissive license, of the Modified BSD or MIT
variety, but I'd like for the copyright notices in each file to be
protected without having to include the whole license itself, if it's
more than a few lines.
The Modified BSD, for example, only protects "the above copyright
notice, this list of conditions and the following disclaimer", so a
simple notice that says "Licensed under the Modified BSD License"
wouldn't be protected unless I included the whole license in each
file. Therefore it's not sufficient.
I think the WTFPL is exactly what you are looking for. It meets all your
requirements, is DFSG-free, and several Debian packages are already
using it.
--
.''`.
: :' : We are debian.org. Lower your prices, surrender your code.
`. `' We will add your hardware and software distinctiveness to
`- our own. Resistance is futile.
John Halton
2008-01-02 16:19:31 UTC
Permalink
Post by Josselin Mouette
I think the WTFPL is exactly what you are looking for. It meets all your
requirements, is DFSG-free, and several Debian packages are already
using it.
Heh. I'd not come across that one. (Text is at
http://www.answers.com/topic/wtfpl).

I notice WTFPL is currently at version 2. Are there plans for a
version 3, being ideally five times as long and introducing new and
incomprehensible (but internationally-applicable) terminology in place
of the word "f---"? :-/

John
Loading...