Justin Erenkrantz wrote:
On 4/21/06, Martin Cooper <[EMAIL PROTECTED]> wrote:
You are correct. Only the years in which the file was actually changed
should be listed in the copyright.
If we want to get pedantic, it should only be year of first publication. ;-)
Yes. What is publication? When we release a major update of the code, that's
a new publication (e.g. if the work is rewritten, as 2.0 was from 1.3, then it's
a single copyright date of the new 2.0 work.)
If we adopt the 'publication' of an ASF work as it's release (collective) than
this whole mumbo jumbo goes away. The date is the release date of the work,
e.g. the date we released 2.0 GA. For 2.2? I don't know if that's a work or
an incremental change to 2.0.
It would certainly save much grief only to coordinate copyright when the work
is tagged from trunk/ to a branch/ and released the first time.
For reference, here's Larry Rosen's post to legal-discuss@ on the form
of copyright notice:
http://mail-archives.apache.org/mod_mbox/www-legal-discuss/200503.mbox/[EMAIL
PROTECTED]
Which contradicts the implemetation adopted by Jackrabbit (recently endorsed
as our model?) We assert our license in every file, yet Larry's post indicates
that is overkill. We fail to assert our copyright in the Jackrabbit src/ tree.
(Although Larry's post suggests a copyright on the web site download page
is enough - which gets interesting on copyright dates since what date would you
brand the page http://www.apache.org/dist/httpd/ ?). Larry is clear that it's
the copyright that binds the license, not the license that binds the copyright,
so asserting a license where you don't assert a copyright seems oddish. E.g.
it doesn't matter if we give you one license, or 100, if we've told you the file
is copyrighted - you can only use it with our permission (the license.)
