Re: Request for a license change

From: SM <sm_at_resistor.net>
Date: Thu, 18 Mar 2010 19:09:07 -0700

At 12:12 17-03-10, Murray S. Kucherawy wrote:

>---------- Forwarded message ----------
>Date: Wed, 17 Mar 2010 14:20:32 -0400
>From: Tom "spot" Callaway <tcallawa_at_redhat.com>

[snip]

>In addition, the clause itself presents a painful burden on distributors
>such as Red Hat, wherein, we would have to take extreme care to ensure
>that if we happen to mention a feature that a component is in any way
>responsible for in "advertising materials", that the "advertising
>materials" include the explicit acknowledgement text. I can tell you
>from Red Hat's perspective, this actively motivates us to not include
>BSD with advertising code in our commercial offerings such as RHEL, not
>because we do not wish to give appropriate credit, but because the task
>of tracking and auditing all possible scenarios in which this clause may
>come into effect is extremely time-consuming and painful, so much so
>that it almost always outweighs the benefits provided by the software!

I do not believe in making things more difficult for distributors using GPL.

>B) In your code, the BSD with advertising license is reworded so that in
>the advertising clause, the word "must" is replaced with "should". This
>makes it a suggestion, and eliminates the additional restriction
>conflict with the GPL. (It does also mean that people would not be
>required to use the advertising clause, but suggests that you would like
>it if they did.)

That's one possible option.

>C) You dual-license the code with something that is GPL compatible,
>perhaps with the GPL itself. If you chose to do this, it would mean that
>it would not be practical for most proprietary solutions to leverage
>your code under the GPL, and they would be forced to choose the BSD with
>advertising license, and you'd have them in the position you do today
>(with all of the concerns addressed above), but for Linux distributors
>and other FOSS projects wishing to leverage your library, they could do
>so under the GPL and avoid compatibility concerns.

I prefer to avoid dual licensing. It is better if we can interpret
the license instead of asking some outside body to tell us what our intent is.

At 12:47 17-03-10, Stefan Schulze Frielinghaus wrote:
>Of course this is true. But I'm concerned with another problem. As far
>as I know a shared or static library licensed as BSD-with-advertisement
>(like libopendkim) is not allowed to be linked into an application
>licensed as e.g. GPLv2/v3. This is one of the main headaches I have with
>this. I would like to release an application using the OpenDKIM library
>but I can't because it is actually licensed as GPLv2.

I would like you to be able to link libopendkim against code
distributed under a GPLv2 license.

>I fully understand this and this is partially my attitude too. I want to
>get credit for the work I've done. So if a company uses my code and
>makes money with it I want to get credit for it. Maybe dual licensing as
>GPLv2 + BSD-with-advertising would be an option?

That's an alternative. But why complicate matters.

>A GPL compatible, more relaxed license my also encourage developer in
>joining the project. At least it would encourage me ;-)

That is an important point. I think that we should encourage people
to join the project.

At 12:59 17-03-10, Mike Markley wrote:
>Debian does distribute things with the 4-clause BSD license, largely
>with the pragmatic view that, at least in the case of BSD code, the
>clause has been deprecated. It's easy to make arguments about the
>requirements of the advertising clause that approach reductio ad
>absurdum (for example: does that mean anyone who prints a pretty Debian
>DVD with a jewel case package and ships them for $5 to cover costs must
>reference OpenDKIM on the cover?), but the prevailing view seems to be
>that it's not an actual problem.

I expected more resistance from Debian. :-) The intent of the clause
is that if you mention OpenDKIM on the cover, you acknowledge the
project. If you sell a DVD with OpenDKIM inside, you don't have to
say so if you are not mentioning its features.

>With that said, I don't see the value in it either--if you want to
>enforce it to increase visibility then it might open up some strange
>arguments, and I happen to think that OpenDKIM will gain plenty of
>visibility just by being distributed, as people discover that DKIM-
>Milter isn't being maintained/updated. But the amount of sweat I've put
>into this is minimal, so don't consider this a push. :)

I am against enforcing the license unless someone does really dumb
things. As I mentioned to some of you, forcing people to do things
can have adverse effects.

At 13:11 17-03-10, Jose-Marcio Martins da Cruz wrote:
>I'm the author of some "free" software and sometimes I'm being
>critisized by some "open source frees software purists". My software
>is free (no strings attached), but the only thing I ask is if it's
>redistributed or used to provide services, that it should be clearly
>stated and advertised. I never asked someone to pay for it, I just
>have a book list at amazon, but in three or four years, just a
>single user bought and sent me a book.

I have released some software under different licenses. I prefer the
no strings attached approach when it is open source. I know that
there are people that can circumvent the intent of the
license. Harassing them about that won't convince them to change.

>IMHO, the main, primary goal of open source free software was to
>create an alternative to proprietary software who were (and still
>are) sold at very high prices and maintaining the user tied and
>dependent. Other reasons come after this idea.

One of the benefits of open source is that you can fix the code if
you find a bug. It's also educational.

>It seems to me that it's a basic ethical and moral duty to
>acknowledge the work of someone if you use it or if you include it,
>without any fee, in your products

Yes but in practice it is rarely done.

>This is even more true for companies such as RedHat, which package
>and sells the work made for other people.

No comment. :-)

At 16:18 17-03-10, Daniel Black wrote:
>The ability to be distributed without pain to the distributor will cause a
>greater benefit to OpenDKIM than I expect the advertising clause ever will.

Yes, the wider the distribution, the more advertizing you get.

>Having a non-license note that we'd like you to advertise this on your
>products would be more favourable here.

Nobody will follow that unfortunately.

The questions are as follows:

  1. Do we want commercial and non-commercial use of libopendkim?

  2. Do we want commercial and non-commercial use of OpenDKIM?

Note that other people will make money out of your efforts. Either
you accept that or you put in restrictions to get them to pay you.

I joined this project because it is code I can use and it is code
others may also find useful. Other people may have made changes that
they don't "give back" to the project. That only affects the project
in the short term. People use the code because it is
free. Developers will use the code if it is compatible with their
pet license. I am in favor of a BSD-like license that the people on
this list find acceptable. Please do not point me to what some
external body says as I will not be convinced by their definitions.

The open approach has benefited OpenDKIM. Although adoption is slow,
there is an increase in the user-base. I suggest that we avoid the
license evangelism trap. Let's stick to making the code better. The
best way, in my opinion, to do that is to consider a 3-clause or
2-clause BSD license.

Regards,
-sm
Received on Fri Mar 19 2010 - 02:09:32 PST