Thomas Bushnell, BSG <[EMAIL PROTECTED]>: > Microsoft takes a bunch away, with the case of qmail, you *didn't* > purchase anything, and you have no rights to copy *anything*--to even > *get* the first copy--except under the terms of the license.
Do all download sites force you to read the licence before you download the tarball? If not, I would guess that by default you do have the right to download the tarball and do what you want with it in private. Arguing from common sense here, consider the case of someone who knows C but doesn't know English. It would seem very unfair for them to be punished merely for downloading the tar ball, editing the code, compiling it and running it. > The purchaser of a copy gets certain rights, and can't have those > restricted without a real contract--that's Dan's argument. But it > doesn't apply to something where you *didn't* purchase anything and > you *don't* have those rights. Are you saying that payment changes the situation here? Consider these cases: 1) I download a file from an ftp site without being asked to agree to anything first. 2) Someone puts a CD through my door or hands one to me at random on the street. 3) I buy a CD. Are you saying I have the right to modify the software in case 3, but not in case 1? If so, what about case 2? Russ Allbery <[EMAIL PROTECTED]>: > The remainder of the page discusses the requirements for distribution of > modified versions. Distributing the tarball, patches, and a script to > compile them is not distribution of a modified version. A court might consider it to be just that in effect. It's like the frequently discussed case of someone distributing a script to compile and link GPL code with GPL-incompatible code. It's possible to argue quite convincingly on both sides of that discussion. > On <http://cr.yp.to/softwarelaw.html>, Dan states explicitly: > > | Note that, since it's not copyright infringement for you to apply a > | patch, it's also not copyright infringement for someone to give you a > | patch. If this is the copyright holder speaking, then this looks like a clarification of the licence, so in the case of qmail he has probably made it true. However, I'm not sure whether it's true in general. Thomas Bushnell, BSG <[EMAIL PROTECTED]>: > I would note that patch files for software are generally thought to be > derivative works, and so I don't see any basis for his claim that they > are not. If I make a list of errata for someone else's book, am I infringing? I hope not. I agree that if the patch is part of a script whose purpose is to create modified binaries of a program whose author has specifically disallowed modified binaries then that is potentially a problem, but if the patch file is distributed for research purposes I would hope that would be all right. Edmund
