Re: What do you win by moving things to non-free?
On Fri, Apr 15, 2005 at 08:48:22PM -0700, email@example.com wrote:
> Apologies if I misedited your e-mail. I am not trying to parrot or
> straw man what you said. It's just complicated and I am trying to make
> my reply readable. If I messed something up, let me know and I'll
> correct my error(s).
A couple asides: Please quote properly; follow the example of the others
on the list. It'll make your replies much easier to read. Also, please
set a name in your From: header; I can't mentally associate something as
silly-looking as "foo_bar_baz_boo-deb" as a human being, and almost
everyone on these lists has a From: header with a real name that at
least looks reasonable.
> I agree that there is a good reason for non-free. I believe the
> justification / reason / purpose of non-free is to earmark data
> (software, etc.) which cannot be redistributed by CD vendors. Java is a
> good example because Sun will not let just anybody hand out copies of
> different parts of it.
No, that is not the purpose of non-free (though it's one useful secondary
purpose). The purpose is being able to package things separately
which are not DFSG-free. Things which are DFSG-unfree but are still
freely distributable are intended--and have always been intended--to go
to non-free. (There have been various discussions of how to further
split non-free, to tag which pieces can be distributed by CD vendors
and which can't.)
The boundary between main and non-free is very straightforward: whether
the work is DFSG-free or not. Nothing else enters into it.
> <<The (exact!) same argument has been applied to lots of useful
> software, such as Qmail, several important Java implementations,
> BitKeeper, and (some years back) Netscape. The argument is
> consistently and resoundingly rejected for those works. Why should it
> be accepted for documentation?>>
> Again, I think the argument (I am not saying I agree/disagree with it,
> I'm just trying to see that it gets communicated) is that the
> non-freeness of other things in non-free is different from the kinds of
> non-freeness existant in GFDL-covered materials.
Qmail is "freely distributable" (if I remember correctly), with
significant limitations on modification. GFDL documents are "freely
distributable", with significant limitations on modification. How
are they so different that they should be treated in any way differently?
> Using non-free should be doable, but not to the extent it becomes a
> knee-jerk reflex or default course of action. If we decide GFDL is not
> free enough, then (this will sound dumb and reduntant to everyone but I
> need to say it) we should not go about sneaking it into main for
Using non-free (or not packaging at all, depending) *should* be the default
course of action, when non-free works are discovered. The GFDL is not free
enough, and not even remotely close: there are non-essential (non-license-
text) pieces that *can't be modified at all*, or even removed. (I find it
continually disappointing that people will actually argue that completely
invariant, untouchable text is "free enough"; I have to wonder why they're