Re: license survey results

[Rich Felker <dalias@aerifal.cx>]

On Wed, Feb 29, 2012 at 06:31:21PM -0800, Isaac Dunham wrote:
> On Mon, 27 Feb 2012 20:55:57 -0500
> Rich Felker <dalias@aerifal.cx> wrote:
> <snip>
> > non-copyleft (MIT/BSD/etc.) crowd:
> > chneukirchen
> > solar
> > nathan mcsween
> > hiltjo
> > khm
> > [rob landley] (in absentia ;-)
> ...
> > LGPL w/static-linking exceptions [almost-]crowd:
> > isaac dunham
> > gs
> > 
> I was thinking "LGPL is semi-bearable, non-copyleft is better, and 
> static link exceptions are a good-enough compromise for now; I don't 
> have any code that gives me a reason to expect you to change to non-
> copyleft."
> So if you want, you can put me under non-copyleft...which is ~7 vs 1 
> for each of four options.

Thanks for clarifying.

> <snip>
> > In summary, it looks like everyone except Luka (and perhaps aep) who
> > responded would like to see at least *some* additional level of
> > permissiveness to musl's license terms, and the largest single group
> > is in favor of non-copyleft/"permissive" terms. As such I'll
> > definitely be making some licensing changes down the line. Please give
> > me some time to weigh the benefits of the different options and focus
> > on the code, especially at this time while widespread deployment is
> > still a ways off. 
> 
> For the record, I'd prefer a license like MIT/BSD to Apache 2.0 & co.
> (where there's a patent clause)--patent clauses just don't sit so well 
> with me. Not sure that it legally makes sense though!

I like patent clauses in principle, but for any kind of copyleft or
similar treatment to be useful, you need leverage. Copyleft, or
respectively the requirement to freely license patents in derivative
works, doesn't get you any code (or respectively, patents) if nobody's
making derivative works... There's also the issue of existing LGPL -
any license that puts further restrictions beyond the LGPL is
undesirable to me because in order to enforce those restrictions we'd
have to make it so new releases aren't also licensable under the LGPL.
Otherwise a party that wanted to avoid the patent clause could just
choose to use LGPL instead.

Rich