>  On Apr 10, 3:34 pm, "Chris DiBona" <cdib...@google.com> wrote:
>  > Basically the answer is when I, Fitz, Greg or the team think it is
>  > popular enough. I know you guys think we don't like it for nefarious
>  > reasons, but what you're missing is we dislike -all- new licenses that
>  > are unpopular. They lead to bifurcation of the open source development
>  > world and that is a high price to pay.

>  Chris, please don't attribute motivations to me that you don't have
>  evidence for.

>  Generally speaking, I actually agree with your position, except that
>  license proliferation per se isn't really the larger part of the
>  problem, which is the combinatorial complexity of license
>  compatibility. Of course, to a first approximation, license
>  proliferation is a pretty good proxy for this, given that many
>  licenses aren't written with maximum compatibility in mind. Also,
>  license proliferation does directly introduce confusion into the
>  license selection process, but that's a relatively minor effect.

>  > I personally think the AGPL is deeply flawed, and I've commented on
>  > that on my own blog and on others, but that really -doesn't- matter.

>  I have my own problems with the AGPL, but for what I want to
>  accomplish, it's the best license available. However, that really
>  doesn't matter either.

>  > If the AGPL gets to be popular, like lgpl or bsd popular, than we'll
>  > certainly offer it as an option on code.google.com, but until then,
>  > it'll be a judgment call on our part. One you might not agree on, but
>  > that's okay.

>  Umm. Actually, I'm fine with it being a judgement call. But I'd at
>  least like some clarity as to at what point you'd *start* considering
>  it. I don't think leaving the lower bound of consideration undefined
>  is going to be helpful to you, but that is your privilege. You will
>  get increasing numbers of "are we there yet?" queries over time,
>  though.

>  However, given the definition you just gave for a 'certainly'
>  boundary, here is my attempt to put a solid number on it:

>  There are 19k projects in SourceForge WWW/HTTP category. Of these, 13k
>  are GPL, 1.3k are BSD, and 2k are LGPL. Four are AGPL (but 15 out of
>  all projects). However, I am also aware of three GNU AGPL projects
>  that are inside SF marked as "Other/Proprietary License", and there
>  may be others.

>  Meanwhile, on Google Code, I see 72 projects labeled GPL, 35 BSD, and
>  16 LGPL. This doesn't seem right, so I use general search instead, and
>  get 351 GPL, 189 BSD, and 92 LGPL. Still seems low. How many hosted
>  projects are there, anyway?

>  Another issue is that few of the new projects that I find interesting
>  use SF for project hosting. Instead, these projects tend to get
>  started on Google Code (which forbids AGPL) or some combination of
>  Trac and Subversion (frequently self-hosted). I don't have a way of
>  doing the same kind of GPL/LGPL/BSD licensing census on those.

>  What I *can* do, is try to compile a list of all AGPL projects I can
>  find.

>  So, here's the question: How many AGPL projects in the wild (which
>  will typically be new web app projects) would meet your "lgpl or bsd
>  popular" standard?

>  - Michael