On Tue, 7 Nov 2006 13:52:27 +0100 Arnoud Engelfriet wrote: > Markus Laire wrote: > > What if a person downloads a GPLed binary and then modifies that > > binary directly? > > If that person can truly say he prefers to hack binaries over > C code, then yes he can distribute just the binary. That makes > sense because there *is no other form* of the software in that > modified state. Exactly, apart from other forms that can be generated from the binary (by disassembling or decompiling) which however are not preferred from his point of view (since he actually use the binary form to make modifications). > > > If yes, then GPL seems to have a severe loophere here, as anyone who > > wants to distribute GPLed binaries without the original source can > > just do some (trivial) modifications to the binary, and then > > distribute this binary as the source. > > Not really. Just changing one static string in the binary is > not sufficient to show that you prefer to modify binaries over > source code. It's very unusual to hack binaries as compared to > source, so a judge (if it ever came to that) would want to see > some serious arguments why you think it's preferred. Indeed, and it is what I meant: I should have written "modify a work in a non trivial way". Thanks to Arnoud for clarifying what I should have said more explicitly (and apologies for not being clear enough in my previous message...) > > If you are a hardcore assembly hacker or you have a very constrained > platform, then it's reasonable to say you prefer to work on binaries. > > > An additional question to consider (in addition to "preferred by > > whom") might be "preferred for what reason". Quite many people might > > prefer to modify the actual binary if it allows them to distribute > > the binary as the source. > > A judge would look at the intent of the licensor when he > wrote (adopted) that clause. The intent of the GPL is to make > meaningful source available. So "I wanted to avoid publishing > useful source" is not going to fly. But "on my 4-bit platform > with 16kB of memory I need to hack the binary a lot or it won't fit" > would be a reasonable argument, and then there's no need to > provide C versions of the modifications. [...] Good: this is the kind of explanation that my previous message lacked... Again thanks for clarifying. -- But it is also tradition that times *must* and always do change, my friend. -- from _Coming to America_ ..................................................... Francesco Poli . GnuPG key fpr == C979 F34B 27CE 5CD8 DC12 31B5 78F4 279B DD6D FCF4
Attachment:
pgp_1v4k7rqii.pgp
Description: PGP signature