Re: PHPNuke license

[John Goerzen <jgoerzen@complete.org>]

On Wed, Mar 05, 2003 at 12:16:23PM -0500, Branden Robinson wrote:
> > In a nutshell, I don't know of any reasonable person that would define
> > "object code" as the output of tr a-z A-Z on a text file.
> 
> Nice to meet you.  :)  That is, I'm perfectly willing to accept that as
> an example of "object code" if the only alternative is to call it
> "source code".

But what I was trying to get at is, outside the context of the GPL, would
anyone writing a general definition of object code consider the output of
"tr a-z A-Z" to fit it?  I think that the answer is pretty clearly "no".

I grant your opinion above is correct, but I don't think it's quite relevant
to the point I'm trying to make.

What I'm trying to say is that "object code" has a certain well-known
definition outside the context of the GPL.  The GPL does not define what
object code is.  Therefore, legally speaking, the normal and conventional
definition is what applies.

Now, an author could explicitly state that "for the purpose of this license,
object code means the output of tr".  That, I think, would be a lot more
reasonable.  I believe the author of some fonts did just that recently.

But lacking such a statement, I don't think that a court -- expert witness
or not -- would find that tr output is included in the common definition of
object code.

> > If a court looks at this, and sees "object code", can we really know in
> > advance if they would use the normal definition or this "liberal" one?  I
> > suspect they would use the normal one, which is another problem.
> 
> What if we had a license like the GPL that used "source form" instead of
> "source code", "transformed form" instead of "object code" and
> "executable form", and "Work" instead of "Program"?

That would make all the difference, I think.

Of course, to the people pondering that change, the ramifications of the
more generalized term should be carefully considered.

> > If the license iteself defined object form that way, that'd be one thing. 
> > (It'd be confusing, but we could evaluate it only one way.)
> > But it doesn't define "object code" at all.
> 
> The FSF does provide a hint, by saying "object code or executable form"
> in two places.  They probably figured an expert witness or two would be
> able to dispose of the issue should it ever reach court.

True, but my output of tr is neither object nor executable, so I don't think
it helps with this particular example.

-- John