Re: Rsync over ssh (pulling from Cygwin to Linux) stalls..

[Christopher Faylor <cgf-no-personal-reply-please@cygwin.com>]

On Wed, Aug 16, 2006 at 10:52:49AM -0500, mwoehlke wrote:
>Christopher Faylor wrote:
>>On Wed, Aug 16, 2006 at 09:22:52AM -0400, Williams, Gerald S (Jerry) wrote:
>>>mwoehlke wrote:
>>>>So...  are we just disagreeing over "safe", or are you actually telling
>>>>me that RH (and thus Cygwin) would *refuse* to incorporate public
>>>>domain code?
>>>In fact that is the case, and it is a shame.
>>>
>>>In RH's defense, there is some legal vagueness around public domain
>>>code (and a few strange laws that add to the confusion by trying to
>>>accommodate shareware and such).  But one would think that as a huge
>>>proponent of open source development, RH lawyers would have figured it
>>>out by now.
>>
>>What Red Hat lawyers (and I, the IANAL, for that matter) have figured
>>out is that if someone sticks a "Public Domain" tag on a piece of
>>software, you still have to go through due diligence to find out if the
>>software is actually encumbered or not since anyone with a text editor
>>can put anything they like in a file.  That would mean getting a release
>>from the person's company.
>
>Ok, *that* actually makes sense. However, that /should/ just mean that 
>they need proof (from whoever would sign an assignment) that the code is 
>public domain, which means it could still *be* public domain, with all 
>the protections (such as they are) that implies.

How would such a form differ from what is currently being used?

>At any rate, I am less confused now; thanks. :-)

I've explained this many times.  At least one in this list.  I think the
last time even resulted in a "zinger" from the "you're off-topic" bot.

>>So, I don't see how public domain buys you much.
>>
>>Oddly enough the FSF has the same requirements.
>>
>>>>The irony of course is that the availability of a commercial license
>>>>makes it look like Daryl's fears are in fact very well founded.  :-)
>>>I don't care if they want to make money, but their current policy
>>>actually prevents code from being released into the public domain.
>>>Even if their real motivation is not monetary, that is a consequence
>>>that I'd rather avoid.
>>
>>Again, Red Hat isn't doing anything different than the FSF.
>
>...Except offering commercial, proprietary licenses. :-) I'm sorry, but 
>it still sounds to me like Daryl is right; signing an assignment to RH 
>is giving them carte blanche to do whatever they want with the code. 

Of course it is.  That's exactly what they're doing.  That's the whole
point of the assignment.

Daryl's contribution wasn't going to cure cancer or cause the lame to
walk.  It was going to fix a problem in Cygwin which probably hadn't
even been noticed by Red Hat customers.

And, of course, the road that Daryl is currently walking seems to require
LOTS of help from people in the cygwin mailing lists.  Were they going to
be cited in his oft-promised-but-still-missing patch?  Should they all feel
hurt because Red Hat would be making money from their efforts, too?

I could see some consternation if major new functionality was being
offered but that is not the case here.  We're talking about a bug fix
which would benefit a lot of people.  Theorizing that Red Hat will benefit
in some tangible way from this kind of patch is an assumption based on
almost no facts.

>The upshot of which is that you can't contribute unless you're willing 
>to let RH do whatever they want with your code. If you're willing to let 
>ANYONE do anything they want with your code (i.e. a permissive license 
>or public domain), then I still think there is a problem if RH won't 
>accept such code (with the acknowledgment that CGF is correct about due 
>diligence; the licensing has to be sound). Otherwise, once RH releases a 
>GPL'd instance of your code, it is GPL'd forever and there is nothing 
>wrong with the assignment (except the original statement; no matter what 
>you have to give RH permission to do anything).

IANAL, but, IMO, "Your code" in this case is not actually "your code".
You're talking about code which modifies Cygwin.  To make your change
you have to study Cygwin's source code.  When you do that, you are
almost guaranteed that anything you do will be GPLed since you have
basically "tainted" yourself with the GPL.

>>Red Hat does not have any interest in the "public domain".  Where Cygwin
>>is concerned, they need to be absolutely certain that they either own
>>the code or that the code's license is conformant with their needs.
>
>That would imply that any permissive, GPL-compatible, non-copyleft 
>license would suffice. So, is there precedent for accepting code that 
>has been placed under such a license?

I'll leave that as an exercise for you to figure out yourself.  Neither
Corinna nor I have a lot of patience for the endless discussions of
IANALs who think they have new points to make on this ancient subject.

The situation with Cygwin's license assignment is what it is.  Corinna
can't change it and I can't change it.  If you don't like it, then don't
contribute but please don't spend a lot of time complaining about how
awful it is because you are just making noises at people who have no
power to effect change.

cgf