[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: AGPLv3 Compliance and Debian Users

Steve Langasek wrote:
On Thu, Jul 11, 2013 at 12:53:01PM -0700, Howard Chu wrote:
Sure, but that doesn't make it DFSG free (hint: it's likely not)[1][2]

[1]: The Dissident test
[2]: The Desert Island test

Sure, but #2 is stupid. We didn't say "must send changes back
immediately." Nor would we wish any such thing; if you're in the
middle of making a long series of changes we obviously want to wait
until the changes are completed and have settled down. Otherwise
someone could make a case that the changes should be sent back the
instant they are written, one keystroke at a time, which is

Send changes back in a timely manner. You obtained the software
somehow; therefore at some point in time a distribution channel was
available to you. The next time such channel is available, send your
changes back. If you're stuck on a desert island and die before such
channel reopens, no one's going to sue you.

I'd say #1 is borderline stupid. It is worded such that it only
applies to hiding existence of a system from the government. Fair
enough; I'm not the government.

That's not the point.  The purpose of the Dissident Test is to demonstrate
that distribution channels for software are not necessarily symmetric; it
may be very easy for you to distribute the software, but very
hard/expensive/dangerous for a recipient to distribute their modifications
back to you.  In the specific case of the Dissident Test, the unreasonable
cost of returning the changes upstream - as opposed to distributing them to
whoever you happen to be distributing binaries to (possibly no one) - is
that sending those communications back may give hostile authorities
information you don't want them to have, such as your location, details
about the software you're modifying, or even simply the fact that you're
doing something that you care about encrypting to keep them from prying.
Even if you aren't otherwise doing anything the government disapproves of,
the mere act of sending these changes upstream might get you labelled a spy.

This is still an unreasonable test. Again, it ignores the element of time. Send your changes at your earliest convenience. If the NSA is breathing down your neck, "convenience" might be a long time away, but that's understandable.

This is one example of why Debian says it's ok for a license to require
modifications to be distributed to your downstreams, but not ok to require
those changes be sent to a particular party.  Users should not have to
choose between complying with the license and being safe from their
government; they should be *free* to exercise their rights on the code in
Debian, even when they aren't free in other aspects of their lives that we
don't have control over.

Freedom always has a price. The price of benefiting from free software should be that you help others benefit from it too. Absolving recipients of all such responsibility merely encourages parasites. Progress happens faster when everyone pitches in, there shouldn't be just a few people creating and everyone else tagging along for the ride. Even here http://people.debian.org/~bap/dfsg-faq.html 12.A.k "This freedom is one of the most important driving factors for progress in computing---and we like progress." That sentence is not talking about this particular point but the underlying concept remains - the goal for all of this is to encourage progress, not hinder it. Hoarding improvements to yourself hinders progress for society as a whole.

  -- Howard Chu
  CTO, Symas Corp.           http://www.symas.com
  Director, Highland Sun     http://highlandsun.com/hyc/
  Chief Architect, OpenLDAP  http://www.openldap.org/project/

Reply to: