Jeremy Hankins <[EMAIL PROTECTED]> writes: > The difference is that a gif is a lot richer than binary software, in > the sense of humans being able to do stuff with it.
This is a real difference. So we have two ways of distinguishing between binary and source: Way one says that binaries cannot be reasonably modified, whereas source can. Way two says that source is the preferred form for modification, and binaries are not. For C code vs. machine language, these give the same answer. I would agree that for other situations, including the case we are considering about images, way number one would allow for more things to be "source" than the second way. *But*, and this is, *once again*, my point: the GPL clearly and unequivocally goes for the second way. In that context, there is no doubt that it doesn't matter whether the form given *can* be usefully modified or used or whatever. The GPL's definition is perfectly unambiguous as it is, and, I think, perfectly reasonable for a copyleft. I have not yet decided what I think about the broader case of defining free software. > I can certainly see the argument that under certain circumstances a > gif would be considered a binary and something like a .xcf would be > required source (for copyleft). But I think it's quite a stretch to > say that that's always the case. That's exactly why the phrase > "preferred form" is so important. Some of the boundary cases would > have to be decided on a situational basis; that's not a reason to say > that gifs can't be copyleft unless they have accompanying source. I have never said that a gif can't be copyleft. A gif might be the actual source. Similarly, assembly language is usually not source, but sometimes indeed it is. In some weird cases, both C and assembly might *together* be the source: for example, where the assembly is a human-modified version of what the C compiler produced.
