On Wed, Mar 15, 2000 at 01:32:06PM -0500, William T Wilson wrote: > > That isn't really important ... > That doesn't really matter either ...
He asked for the difference, that's my answer. I did not say it is relevant in copyright law. On the contrary, I said it doesn't in the next paragraph. > > However, note that the output of a program is usually subject to > > copyright law, so a pipe is legal if the license for the output allows > > usage of it. The GPL explicitely does not restrict usage. > > Only if the output of the program is considered a derived work. Where do you get this from? Or, the other way around, what makes you think it is not a derived work, as it is the output of the program? > For > example, a program that prints its source code would have its output > covered by copyright. But a program that outputs arbitrary data does > *not* place its output under the copyright of the original program. Tell me why, please. > If > you use a program to create a separate work, you own the copyright to it. How can it be seperate if you used the program? Only if there are no traces of the program. > If you use a word processor, documents you write with it are yours; I agree as long as the format is simple (for example plain ascii with only the text you wrote in it). > if you > use a compiler, copyright on the executable belongs to the owner of the > source code, The executable contains parts of the compiler, startup code for example. All compilers should (and do, AFAIK), grant you a permissive license on this code. autoconf does give you a license on the generated configure program. Likewise for bison and flex. > and if you use Acrobat Reader to print a PDF, then the > copyright on the results belong to the owner of the PDF, not to Adobe. I have no license of Acrobat Reader (Writer?), so I can't say anything about it. > > explicitely allow it. Piping is restricted, but allowed as "use" by > > all free software licenses (if you don't bundle the application and > > And by copyright law in general, as fair use. Yes, probably. I don't know any case law, though. > > distribute them as a whole). Calling kernel functions is restricted, > > but allowed by the linux kernel license (or the other way round, by > > the GPL in the exception clause). > > Mostly this is a result of having to include some portion of the kernel or > library in the program that wants to include it, as headers. This makes a > derived work. Using a pipe does not create a derived work since the two > programs are separate even while the pipe is being used. This is what you and many people think. However, I would not give my right arm for it. Unfortunately, there is little case law. There is some case law that supports such arguments (nintendo), however, in those cases, both works are not distributed together, and such it is a different matter (usage vs distribution). I'd say that it doesn't matter what the dependency is, linking, piping, or whatever. As long as two tightly bound programs that rely on each other to operate properly are distributed together, they constitute a combined work derived from either of both works. Thanks, Marcus -- `Rhubarb is no Egyptian god.' Debian http://www.debian.org Check Key server Marcus Brinkmann GNU http://www.gnu.org for public PGP Key [EMAIL PROTECTED], [EMAIL PROTECTED] PGP Key ID 36E7CD09 http://homepage.ruhr-uni-bochum.de/Marcus.Brinkmann/ [EMAIL PROTECTED]
