Re: Re: Handling special characters (\/:*?"<>|) gracefully

[Christopher Faylor <cgf-no-personal-reply-please at cygwin dot com>]

On Fri, May 26, 2006 at 01:20:57AM -0400, Joe Smith wrote:
>"Christopher Faylor" <cgf-no-personal-reply-please@cygwin.com> wrote in 
>message 20060525143449.GC12123@trixie.casa.cgf.cx">news:20060525143449.GC12123@trixie.casa.cgf.cx...
>>On Thu, May 25, 2006 at 02:55:24PM +0100, Dave Korn wrote:
>>>On 25 May 2006 14:45, Williams, Gerald S (Jerry) wrote:
>>>>The code was released as PD, so there are no copyrights claimed.  Once
>>>>you get a copy, you can do whatever you want with it.
>>>
>>>Well, I'm not 100% sure that means I can re-license it.  And IIUIC I
>>>would need to actually *own* the copyright on anything that I want to
>>>place under GPL.  So cgf's caution is probably correct and I should
>>>really do a clean-room implementation of my own.
>
>If you modify it in any way, you have created a derived work.  That is
>very well established.  If your modifications are non-trivial, then you
>would hold copyright on the derived work, so there should be absolutly
>no problems releasing the dirived work under the GPL.
>
>There is considerable precident in the publishing industry to back this
>up.

I'm pretty sure that you can't substantially modify someone else's work
and then change the entire license to the work.  You would own the
copyright on the parts that you modified but the original copyright
would still apply to any unchanged parts.  And, I belive that simple
text replacements of variable names, etc., would not make a derivative
work.

>>>I think it'll be ok to use the list of functions that you suggested as
>>>the basis, however!
>>
>>FWIW, Red Hat's legal department (who no longer respond to my email)
>>once told me that incorporating public domain stuff into Cygwin was ok
>>as long as it was very clear that there was no claim on the code
>>
>>Unfortunately, proving that a company doesn't claim ownership is
>>usually roughly equivalent to having them fill out the idious cygwin
>>assignment so, AFAICT, this isn't usually all that useful an option to
>>pursue.
>
>I suspect that if Jerry were to post a scan of a signed document
>indicating that the company does not claim authorship, or that the
>company has placed the work into the public domain, it would be
>difficult for the company to claim they did not.  Now, that should
>really be acompanied by a similar document indicating that Jerry does
>not claim rights to the work, as Jerry could potentially have some
>claim to the work.  For example it is possible that the contract that
>gave the company any claim to the work was never valid in the first
>place.  Highly unlikely, but theoretically possible.
>
>DISCLAMER: IANAL, but I spend a lot of time reading Debian-legal, so I
>have a fairly good understanding of how copyright works.

One of my reasons for posting what I did about Red Hat lawyers was to
make it clear that we won't be getting help there.  They don't respond
to Corinna's email either and she works for the company.  I don't think
they respond to Corinna's *boss's* mail, for that matter.  They just
don't care about Cygwin, apparently.

(although I think I'd love to be proved wrong about this)

So, any amount of opining and IANALing is not going to be very fruitful.
I'm pretty sure that Corinna is not going to just allow changes into
Cygwin just on the say-so of IANALs.

So, while someone did tell me that public domain software might be ok to
include, practically speaking, we only have one approved mechanism for
getting changes into cygwin - the person and the person's company who
makes the changes must have a signed agreement with Red Hat.

cgf