Re: Proposed license policy

[Havoc Pennington <hp redhat com>]

Ian Peters <itp helixcode com> writes: 
> I think it's a great idea to have a dialog about this.  I think we'd
> all like to see GNOME as the development platform of choice for all
> software, free and non-free.  I don't know whether that means we
> should aim for parity, though.
> 

My view on GPL libraries is the following.

Assume that we have a library X that no proprietary platform has an
equivalent for. (GConf is actually such a library, but you can imagine
a more important/impressive one.)

 1) right now our overall platform is significantly behind proprietary
    ones; we just recently caught Motif/CDE, and have a lot to 
    do before we catch Java/.Net/MacOS X. Here I mean entire platform;
    I think we've caught them on the GUI toolkit front, pretty much,
    with GTK 2; but there are lots of remaining gaps. If we have
    a spiffy library X, then we should use it to offset
    our deficiencies in other areas; if it's GPL, then it helps 
    us implement end-user features, but doesn't help us 
    with the platform.

 2) No one is going to free their software just to use this library 
    X. The reasons are:
 
  a) if library X is really necessary to write apps, there will 
     be a proprietary equivalent

  b) Even if there is a proprietary equivalent: if your whole business
     model is licensing fees, you are not going to change that to
     avoid rewriting a library; even a library with 200K lines of code
     like GTK (see Mozilla, StarOffice, Applix - none of those used a
     prebuilt GUI kit, though Applix has switched).  Moreover you can
     easily fund replacing a library with licensing fees. I can't
     imagine a library cool enough that people would want to change
     their business model.

   You might say, "but the GPL has made people release code in the
   past." True, but usually code they weren't selling anyway and of
   the cases I know the GPL was applied to an application or kernel,
   not a library. And there really are not very many of these cases.

 3) I think the way to succeed is to make companies want to free their
    software. Then you get a bunch of momentum and developers from
    real commitments by that company. cf. Eazel for
    example. Abandonware that someone was forced to GPL is going to be
    grotty, useless ex-proprietary code with no maintainer. This is
    not helping us much.

 4) Developer mindshare is very important. The way to get that is to 
    get people using our libraries at their day jobs. For the next 
    few years at least, most of the time those people are writing 
    proprietary code. If GTK and GNOME programming is a marketable 
    skill, we have way more people with an investment in our 
    technology and the ability to submit patches.

 5) Even with free software, some is MPL or the like, and we should 
    probably allow people to port such apps to our platform.

 6) A possible exception to all of this is the case where we might
    have a patent on key parts of the library, such that no
    proprietary platform can duplicate it, and the library is really
    compellingly useful. This might be interesting, but if the library
    is really compelling you may as well write "sue me" on your
    forehead. In any case, we have no such library at the moment.

Havoc