hellekin
2013-04-01 16:45:59 UTC
Erkan Yilmaz: What are the plans of the GNU social team to
continue StatusNet ?
*** I've been asking the same thing or similar to mattl for a while.continue StatusNet ?
Since he considers I wanted to takeover the name of this project, he
ignored me consistently. Last thing he said was that GNU Social is not
dead and they're going through a procedure to incorporate "two huge
code donations". Although he never replied, I suspect that one of them
is the code of StatusNet itself. Gathering copyright ownership on that
one is titan work.
The copyright issue with StatusNet is that there's no single owner: it
prevented Evan Prodromou, the man behind StatusNet and Pump.io, from
being able to propose commercial dual-licensing of the software after
the fact, as it would require cooperation of all copyright owners.
It's not a problem with AGPLv3+ itself, the license of StatusNet, but
of the copyright ownership strategy: the copyright ownership strategy
of GNU Social is to ask all contributors to waive copyright ownership
to the FSF, so that the foundation can defend the software in a court
of law.
When a project has many developers sharing the copyright, vs. a single
entity, changing the license ranges from very difficult to impossible.
For example, DokuWiki is released under the GNU General Public License
version 2 exclusively. At the time of its creation in 2004, it seemed
like the normal thing to do, as GPLv3 did not exist yet. Over the
course of 3 years, until the release of GPLv3, the number of
developers--and copyright owners, had rocketed to more than a hundred
contributors. Getting them, including the missing ones, to agree on
adding the little + to GPLv2 in order to authorize the GPLv3, and
later versions, in addition to the current license, was already a lost
cause. [0]
However I'm not convinced of the strategic interest of such a move--if
it's what I suspect. The successor of StatusNet, pump.io, comes with a
number of advantages--when it's ready: the most important is probably
the performance gain. Evan commented recently [1] that the cost
savings of hosting pump.io vs. StatusNet's costs are more than
ten-fold, which means that pump.io could easily run on a Freedom Box,
a Raspberry Pi, or the equivalent free hardware (aka 100% OHL [2]
compliant box) when it's available, for any household to run
cheaply--server-resource-wise, and be more scalable for smaller
organizations that want to run their own instances.
Other concerns include the viability of the code base in the future:
StatusNet is already mature code, and probably hit the wall of
diminishing returns already, while pump.io is nascent and built on
NodeJS. NodeJS is a much younger platform than PHP, and benefits from
faster development cycles, and an innovative community who learned on
the errors of the past. The current competitive advantages of PHP, its
pervasiveness among ISPs, and its very large developer base, are going
to be challenged very soon, as ISPs adopt newer languages such as
Python/Django, Ruby/Rails, NodeJS, etc. more widely [3]. Note that a
sound copyright ownership policy on StatusNet might revive the
innovation of the software, although it will still remain
(theoretically) slower to develop that NodeJS, Ruby, LISP, etc.
Now, I hope that the GNU Social team will be more responsive as they
approach release date. Mattl has been busy with the very successful
LibreFM, and the daunting secret task of copyright assignments over
the upcoming secret weapons of GNU Social.
Let's hope for the best.
==
hk
[0] Dokuwiki's License FAQ does not mention any specific version of
the GPL, while the repository explicitly mentions GPLv2, and the Ohloh
entry mistakenly mentions GPLv2+ (edit: corrected)
https://www.dokuwiki.org/faq:license
https://github.com/splitbrain/dokuwiki/blob/master/COPYING
http://www.ohloh.net/p/dokuwiki/licenses
Dokuwiki's wiki contents though, went through a license change 2 years
ago:
https://www.dokuwiki.org/licensing_change
https://www.dokuwiki.org/devel:ideas:relicensing
[1] http://identi.ca/conversation/99265765#notice-100338214
[2] http://www.ohwr.org/projects/cernohl
[3] http://networkeffectalliance.org