afl 2.1 license

[Havoc Pennington]

On Tue, 2004-11-30 at 14:04 +0530, Michael Meeks wrote:
> 	I guess it is better - but I guess we would have to do some painful
> legal work in-house to review the terms. At least the immediately
> obvious, no-brainer "this will never fly whatsoever", poison pill
> appears to have gone (to my untrained eye). Of course, what a
> combination is I have no idea.

It has specific legal meaning I believe. I don't think your lawyers will
choke on AFL - sort of the whole point of AFL is that it's a lawyer-
friendly X/BSD type of license.

The new weak patent clause as I understand it is simply that you can't
license a specific version of dbus code, then sue because that specific
version infringes a patent, or you lose the license to dbus. That seems
completely reasonable to me. In other words to use dbus you license your
patents *only for use with dbus* *until such time as you stop using
dbus*. You can't be using the library and simultaneously keeping others
from doing so with patents.

Disclaimer, the original license text is the place to look for the real
meaning of the clause, I'm just trying to interpret it.

> 	So - I'm still concerned that we throw away a license that is known &
> accepted, for one that is new and unproven for no gain I can easily
> understand. I thought we were going to switch to an LGPL/AFL combo ?

Doesn't NLD already have HAL/dbus in it? RHEL is going out with it.

The value is pretty simple; AFL's patent clause and general legal
clarity are nicer than the incredibly vague and confusing LGPL.
In addition to the LGPL's inherent suckage, the LGPL option renders the
AFL patent clause useless since there's no reason to accept the AFL
license.

Havoc