Re: [autonomo.us] open web v. standardized web

[Bradley M. Kuhn]

Luis Villa wrote at 23:52 (EDT) on Wednesday:

> It's not GPL-forcing-you-to-open-your-code-to-play-in-the-sandbox
> open, but even RMS hasn't argued for that at the platform level in
> many years, at least until the hypothetical day when the underlying
> platform is so successful that you actually have the leverage to make
> that work.

A number of people have asked me about the idea of an Affero LGPL in
this context: something to make sure that the platform doesn't just stay
open, but also free as in freedom, even if some of the applications
aren't fitting with the Franklin Street Statement and the usual four
freedoms.

The "Affero LGPL" idea has some merit, in the same the idea of making
Gtk and other GNOME libraries LGPL did: proprietary applications on a
Free Software infrastructure that hackers can modify will allow more
freedom to users than the alternative.

This also relates to a proposal that Asheesh made at the autonomo.us
track at LibrePlanet back in -- IIRC -- March 2009.  The idea is to
create hacks on either side of proprietary services: make a new Gmail
interface by intercepting the Gmail API, and also make your own backend
by doing the same, and they can be completely separate projects, which
maybe could get merged once they both worked really well.

Hacks like this are important: we need to encourage smaller tasks that
allow at least hacker-ish type people to migrate away from proprietary
network services somewhat now, even if they don't fully give them up
entirely all at once, while they wait for the Free Software project to
"finish".

It's also not horrible to take currently proprietary protocols that we
can successfully reimplement as de-facto standards [0].  Again, I make
the analogy to GNU: Unix was a proprietary de-facto standard when GNU
started (POSIX didn't exist yet).  Maybe Gmail's API is the proprietary
de-facto standard for email in the net.services era.  I doubt Google
would sue us if we reimplemented in a non-copyright-infringing way, and
it's really low hanging fruit, which was the center of Asheesh's point
at the time.


Meanwhile, I think this Affero LGPL idea is somehow related to Asheesh's
idea: maybe we need some of these API libraries under this "Affero LGPL"
so you can mix-and-match for the moment with proprietary stuff like
Gmail while we work on the full-fledged replacements.  What do others
think?

[0] I'd note that Skype is probably a counter-example here.
    Reimplementing around Skype is such a massive undertaking, and it's
    so difficult to reverse engineer what they're doing that we just
    have to go from ground-up to replace Skype.  I'm hoping the "no NAT
    in IPv6" thing will help SIP not be so damned difficult and all the
    freedom-respecting VoIP solutions available won't require massive
    technical knowledge to make work.
-- 
   -- bkuhn