Here's my promised message about the rsync discussions in and outside the 
working group.

The sidr mailing list shows exchanges about rsync around Oct 30-31, 2006 
with Joe Abley expressing concerns about interop of different 
implementations, and 28-31 Mar 07 starting with William Leibzon expressing 
concerns about IESG/IETF process.  Those message also include messages 
describing the features that are needed and how rsync satisfies those.

That's about it.

But the discussion during meetings and in private exchnages has 
snowballed:


(1) rsync is not a URI

OK, we'll get it registered based on the RFC

(2) rsync is not an rfc and the author is not interested in making it one

OK, we'll do a third party registration based on the spec

(3) there is no stable spec for rsync (aka "the code is the spec")

OK, we'll register the URI based on ... the code? (if we can)

(4) there are (a few) wg concerns about non-interoperability of 
independent implementations, due to lack of spec

OK, we'll presume that everyone uses the open-source implementation.

(5) rsync has gone through some non-backward-compatible version changes in 
the past.

OK, we'll hope the rsync project doesn't do that again, else new versions 
of rsync with non-backward compatible features would require a flag day.

You may recall a mention in Monday's meeting of license difficulties for 
commercial implementations.

At this point in the history, I discovered that the rsync open source 
project used a license that forbid charging for any code that included 
their code.  That would have made commercial implementations problematic.

However, I just discovered that with the announcement of rsync 3.0 on 
March 1, 2008 they have moved from GPL version 2 to GPL version 3:
http://www.samba.org/rsync/GPL.html

>From a first glance through, I can not find similar text that would 
prohibit charging for code that includes their code.  People who are used 
to reading these GPL licences should take a look to see if you see limits 
on commercial use of the rsync code.

So the difficulty about commercial implementations could be OBE.

--Sandy
_______________________________________________
Sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr

Reply via email to