Anthony Towns <[email protected]> writes: > When I take, say, /bin/ls, and run it under fakeroot (and thus link it > with libfakeroot.so.0, aiui), I'm not (afaik) causing the resulting modified > file (that's only ever in memory) to carry prominent notice that I changed > the files, and I've no idea if the date of the change is stored anywhere.
I think Raul's examples have gotten a little carried away here. :) The GPL here is speaking of source files, and you haven't modified those; it also explicitly limits its scope to acts of distribution. If no distribution is taking place, then the GPL lets you copy with impunity. (Mind you, internal sending of copies from one unit of a large corporation to another unit might well constitution distribution.) Note carefully that I did *not* say "if you aren't distributing anything, then you aren't violating the GPL". Whether there is distribution is a question that is settled by looking at the total acts of the relevant group of actors, and not just at one person's single action. The problem with "user-does-the-link" subterfuges is not that linking is itself the creation of a derivative work. (It is the creation of a derivative work, but that copying isn't the problem.) The problem with "user-does-the-link" is precisely in the fact that it's a subterfuge. The law doesn't in general care how you structure a single complex act or how many people you divide it among. Each willing participant in a big group action is just as liable for the act of the group as if they had done the whole thing themselves. (This sounds like conspiracy law, but it's actually a quite different. I'm happy to explain what the difference is if you like.) Thomas
