Anthony DeRobertis <[EMAIL PROTECTED]> writes:
> On Mon, 2003-10-13 at 22:01, Brian T. Sniffen wrote:
>
>> Let's say Alice's installer uses secret-sharing or error-correcting
>> codes to meld the program and the documentation, then produce separate
>> works from them.
>
> Like tar czf?
Not quite what I had in mind: I was considering something clearly a
program and using, not merely aggregating, the two works: something
which would invoke the FSF's ridiculous assertion that dynamic linking
is modification.
>> Let's say Alice distributes them as an InstallShield(tm) program, or
>> as a shar-style archive: an installer program which installs the
>> documentation and the useful program. Certainly nobody can make such
>> an installer -- which is a derived work -- except Alice.
>
> "which is a derived work" is quite questionable. It'd probably be a
> "mere aggregation" --- certainly just as much as a ext3 filesystem.
So given that, no, I don't mean anything like a tar file or a
filesystem: I mean something more like a closure which returns other works.
> How are tar and shar different, legally? None, I'd bet.
>
> I don't think taking an archive file, and including an unarchiver
> (InstallShield) is any different, either.
This is why I was careful not to describe it as an archive file and an
unarchiver: it is a program which produces output; that output is a
copy of a copyrighted work. There isn't a clear data section; rather,
the useful program (which Alice originally wrote and Bob modified) and
its documentation are organic parts of it. Perhaps it links against
Alice's program and the documentation.
-Brian
--
Brian T. Sniffen [EMAIL PROTECTED]
http://www.evenmere.org/~bts/
