On Sat, 2002-08-31 at 10:18, Edmund GRIMLEY EVANS wrote: > If you took the obfuscated code, did your best to unobfuscate it by > applying both automatic reformating and manual editing, and then made > some functional changes in it, or even non-functional changes, such as > adding comments, I think you could then claim that what you have > created is now the "preferred form for making modifications" to the > new, derived work. Problem solved. > > You could compare this with the case where a program was originally > written in Pascal, but someone automatically converts it into C and > starts making changes in the C version. Then at the least for some > people, the "preferred form for making modifications" would be the C > version and not the Pascal original.
Yes, this is very true. However, I would consider this to be a fork from upstream. It's comparable to taking a binary-only module in a GPLed product, reverse-engineering it, and writing new source that duplicates the module's functionality. What bothers me, though, is that the de-frotzed modules we create would be derived works of the frotzed ones. Since, technically, "the preferred form for modification" was never provided for those modules, there's still some uncertainty. I don't know if upstream could exploit this as a legal loophole in their original license (not in the GPL itself, but in their use of it) and punish the people who did such a thing. That, of course, is one of the reasons why I advocate caution. OTOH, if our de-frotzed modules were licensed under the GPL, and if we claimed there was an implicit exception to the GPL in the upstream source, we could deny upstream access to our changes, since we wouldn't need an implicit exception.
