Open Stack

So this may need to pop off to -legal soon :). And IANAL ... but...

On 23 December 2013 17:35, Chet Burgess <cfb at metacloud.com> wrote:
> It's unclear to me what exactly constitutes writing a new patch. I can check
> out oslo.messaging, and without trying to merge the patch just go and make
> the same change (its literarily a 2 line change). I can write the tests, and
> I can submit it (which I'm happy to do, I really want this bug fixed).
> Honestly though this change is so trivial I don't see how my patch would
> look all that different from the one already posted. I know there is prior
> art. The mixin class that kombu provides does the exact same thing. Is that

Prior art is a patent concept, not related to copyright. Copyright is
entirely about copying. If you haven't read the patch, sit down with
the source code and the bug, and write up a new patch, any resemblance
or lack of resemblance is irrelevant.

If you *have* read the patch, and sit down to do the same thing, it
becomes possible for questions to be raised about whether you copied
the patch or not - particularly if the patch you create does end up
looking the same. But - I don't think OpenStack needs to be a moral
compass here - honour the CLA when you make a contribution, enough
said.

> sufficient? What else would need to be done to make this free an clear for
> our use? I'm going to try reaching out to the author to see if I can sort it
> that way, but this still seems like there is a general problem here.

My advice, if you have concerns:
 - don't read patches on the bug tracker
 - perhaps even delete them - patches should be in Gerrit.

> Given the current interpretation of the IP laws someone has an effective way
> to block progress on a feature, blueprint, or project as a whole. Create a
> launchpad account, don't sign the CLA, just start posting implementations to
> launchpad. If the simple act of reading the bugs now encumbers us from being

I don't think it's that simple.

> able to fix them in a certain way or using certain patterns we have a

Copyright != Patents. Patents describe processes - patterns. Copyright
is on the actual work + various transformations of it. A patch with an
implementation of linked lists would not stop us using linked lists no
matter how many folk read it ;)

> potentially serious issue. If this is really the case should we not lock the
> bug tracker to only those who have signed the CLA or have the TOU clearly
> state that any code posted is automatically ASLv2? Am I misunderstanding the
> scope of this problem?

I think you're over estimating the risk. It's often entirely possible
to read a patch and then write a de novo patch yourself.

-Rob

-- 
Robert Collins <rbtcollins at hp.com>
Distinguished Technologist
HP Converged Cloud