[conspire] The old stealth licence change trick
rick at linuxmafia.com
Wed Sep 29 23:31:58 PDT 2010
And one last comment about the form usually taken by open source to
proprietary transitions: There's almost always a ratcheting-up effect
The current, initial proprietary variant of TACO (now Abine) states that
it's for 'non-commercial uses consistent with the description of the
Products we provided with the Products'. Non-commercial-usage-only
licensing restrictions are almost always the camel's nose in the tent --
and, in cases of well-known software such as SSH, which started out GPL
and then went gradually proprietary when inventor Tatu Ylonen formed a
commercial company, inevitably a bunch of apologists come out and say
'Well, who does that really hurt? Most uses aren't commercial, and you
and I can still lawfully use the program the way we always have, and
nobody will notice if we occasionally cheat and use it at work, too.'
The owner's aim is to get people used to the restrictions gradually, and
add more over time. Once people misplace their master archives of the
last open source version's source code, the owner doesn't have to worry
about emergence of open source forks any more.
This happened, very famously, with BitKeeper's gratis-use version,
during the period when the Linux kernel community was becoming more and
more dependent on it:
o Earliest versions had available source code, was effectively usable
only for non-commercial use on account of public 'Open Logging' of
your data, and had a licence clause that the licence would convert to
GPL if the firm ever shut down its servers for 180 days.
o Then that clause was withdrawn.
o Then a clause was added requiring users to upgrade to any new
version put out. (This meant that you could not evade new
restrictions by not upgrading.)
o Then the source code access was withdrawn.
o The a clause was added permitting the company to cancel licences
deemed to have created too much support costs. (This was later
o Then a no-compete clause was added, cancelling the licence if you
did any work on any other 'substantially similar' version control system.
o Then, 2005-07-01, author Larry McVoy remotely disabled all users'
copies, in reaction to my friend Andrew Tridgell 'probing' BitKeeper
servers by connecting to them using telnet and typing the command
'help', using this knowledge to develop SourcePuller, a simple tool
to extract BitKeeper data with all metadata intact, e.g., for
compatibility with another version-control system. McVoy claimed
this action violated the no-compete clause.
(Torvalds finally reacted to the 2005 events, after years of making
excuses for McVoy, by sitting down and creating git.)
More information about the conspire