[AFS3-std] Copyright, internet-drafts and .xg files

[Jeffrey Hutzelman]

--On Thursday, February 25, 2010 11:54:28 AM -0500 Jeffrey Altman 
<jaltman@secure-endpoints.com> wrote:

> This is clearly not the IBM Public License 1.0 and I do not believe that
> it should have been replaced by it.

Our interpretation at the time was that IBM's release of OpenAFS under the 
IPL superceded other IBM license terms found on individual files within the 
package.  It was clear that licenses on many source files were not cleaned 
up, and that it was IBM's intent to release the package as a whole under 
the IPL.

We replaced IBM copyright notices and license terms on individual files 
with those of the IPL in nearly every case.  We did not remove or change 
non-IBM copyright notices, and we did not replace license terms on software 
like NTP that was distribute with OpenAFS but not as part of it, or on 
files like some of those from Sun that came with less restrictive terms 
that were not IBM's to change.

We believed these changes were necessary before we could redistribute 
OpenAFS in any useful way.  In particular, we believed the IPL did not 
allow us to modify and redistribute OpenAFS _as shipped by IBM_ without 
first insuring that the IPL license language appeared everywhere it was 
supposed to.


Note that just because some file with a less restrictive license appeared 
in the original release does not mean that the same file plus 10 years of 
modifications can still be distributed under that original license.  In the 
intervening 10 years, those files have been part of OpenAFS and subject to 
the terms of the IPL; thus, the only way to use or distribute the modern 
versions is to comply with those terms.




> As for the .xg files, the majority of them did not include any copyright
> statement at all.

As you know, that does not mean no copyright applies.  Again, as those 
files were part of the OpenAFS distribution distributed by IBM, we assumed 
that IBM intended to apply the terms of the IPL 1.0 to them.  In the 
absence of that assumption, we would not have been able to distribute the 
files at all, because we would have lacked a license allowing us to do so. 
Lack of a copyright notice or specific license terms on a file does _not_ 
mean that it is OK to do whatever you want with it; it means you need the 
owner's permission to do anything at all.


> Those that did had a much earlier copyright statement
> dating back to ITC CMU project (1987,1988).

Not relevant.  All of that IPR was transferred to Transarc as part of the 
spinoff, and eventually became entirely the property of IBM when Transarc 
became a wholly-owned subsidiary.  The original ITC files can be 
distributed under the original terms, which are probably fairly generous, 
but the ones that belonged to Transarc and IBM for 10+ years can only be 
distributed with the permission of IBM.


> For starters, Simon, go ahead and publish your I-Ds using the .xg
> sources from the IBM OpenAFS 1.0 release.  They should meet your
> needs.   OpenAFS has a mess that needs to be cleaned up and that will
> need to be done in conjunction with IBM.

Only Simon can decide whether he feels comfortable submitting I-Ds using 
files from any OpenAFS release, including the original IBM contribution. 
Neither you nor I are lawyers, and we should not presume to advise him on 
this issue.

-- Jeff