[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: PHPNuke license

On Tue, Mar 04, 2003 at 12:50:13PM -0500, David Turner wrote:
> I think that's the claim -- that certain modifications of PHPNuke are
> forbidden.

Okay, just reassuring that we're on the same page.

> Interestingly, I don't think (2)(c) would forbid a modified PHPNuke to
> print the copyright notice to a printer (or console) in the server room,
> instead of on the web page the user sees.  The more I look at the
> clause, the more convinced I am that its sole purpose is to torture me.

See, this is exactly my problem: who is it notifying?

If it prints the notice to a printer, it's notifying the site administrator.[1]
If it prints to HTML, it's notifying the users of the website.

The former is reasonable to me, at least as far as the GPL goes.  (However,
PHPNuke and Apache are clearly (I hope) not being started interactively wrt
the site admin, so these notices shouldn't be mandatory.)

The latter is the case I dislike.  I think of website users using the
website (and only very distantly and nebulously are they in any way
users of PHPNuke), and the administrator as the user of PHPNuke.

> > (Incidentally, I've always thought of programs like "gdb" as being
> > interactive, and programs like "df" and "cat"--including ones that
> > happen to read data from stdin--as being uninteractive.  I don't mind
> > debating these interpretations of "interactive" here, but it's tangental
> > to the PHPNuke discussion.)
> Do you say that it's tangential because you think PHPNuke is clearly
> interactive?  Or because, interactive or not, Debian should consider its
> interpretation of (2)(c) non-free (whether or not it's correct from a
> legal perspective)?

I mean that the idea is that PHPNuke (run by Apache) and "sort" (in the
middle of a shell pipeline) are equally interactive.  Neither talks
directly to the user; they go through other programs.  So, if
that argument is bought, then we have two choices: either sort(1)
outputting a mandatory GPL blurb on startup is DFSG-free (gross), or
PHPNuke outputting a mandatory GPL blurb to HTML is DFSG-unfree.  (And,
in parallel, either sort(1) outputting a GPL blurb would be mandatory,
or PHPNuke outputting one would be optional[2].)

Like I said, I havn't been completely convinced of this.

(Determination of whether "df" is interactive is tangental to this,
though I'm not uninterested in reaffirming the notion that it is not.
Determining the interactiveness of "cat" in a pipeline would be
relevant only if we agreed to the above.)

> I actually am inclined (after looking at the Foldoc definition pasted in
> another message in this thread), to agree on df and cat.  It's the
> back-and-forth which defines interactivity.  Interestingly, this
> provides a "out" for Apache other than the two I listed above, since
> HTTP is technically stateless...

Oops.  Once we start using the word "technically", we're getting outside
the realm that licenses should care about ...

[1] ignoring the common case where the system admin and the site admin are 
two different people and the site admin has no physical access to the machine

[2] however, in this case we go back to the WU situation: a copyright
holder's strange interpretation rendering a normally free license non-free.
This would be annoying, since it would probably be the first time a GPL
program was declared DFSG-unfree (and we might hear rumblings from the
hopefully small DFSG#10-as-grandfather-clause crowd) ...

Glenn Maynard

Reply to: