Re: Final text of AGPL v3
On Mon, 19 Nov 2007 23:35:35 +0100, John Halton <firstname.lastname@example.org>
For now, as regards the "interaction" issue my inclination is to
assume that it is implicit in this that (a) we are talking about some
level of direct input/output interaction from the user's point of
view. i.e. software providing user interface-level interaction, not
software further down the stack, and (b) we are talking about
"interaction for the purposes of exercising the intended functionality
of the program", not 403 errors.
That still leaves us hanging wrt. a web-server or network driver that
is under the AGPL3. For a web-server, serving a 403 error is (part of)
its intended functionality, as is sending the packages for the network
driver. Also, the web server is "clearly" interacting with the user
(receiving input originating from the user and responding to it).
The web server might be able to "prominently offer" access to the source,
but the network driver surely can't.
So, the "prominently offer" part suggests that "interaction" means that
the user must *see* output from the program *itself*, not just something it
So an interactive AGPL program *must* have a prominent way of giving
information to the user. This breaks for any server where the return format
is restricted (web servers should not insert content on delivered pages,
so unless the HTTP headers are prominent enough, an AGPL3 web server is in
The problem here is still that "interaction" isn't defined. And probably
can't be without being either too narrow or too broad.
/L 'some categories are best defined by their perimeter, others by their
Lasse R. Nielsen - email@example.com
'Faith without judgement merely degrades the spirit divine'
Reproduction of this message, or parts thereof, is allowed if proper
attribution is given.