On Fri, 2004-07-23 at 23:41, Don Armstrong wrote: > If the pictures are built from XCF files or PSD files instead of being > created tabula rasa as a jpeg or gif or whatever, then the the format > that upstream actually uses for modification or creation of the work > should be supplied. You make a jump here that I, as an upstream author, don't like. I'm perfectly happy providing the preferred form of modification for my works -- but that is not always the same as the preferred form of *creation* of my work. I might make a throw-away shell script to generate some code or data, and then throw it away. From that point onwards I'll edit the data by hand. Or I'll grab an image off my camera as TIFF, but only work with (i.e. edit) it as a JPEG. Or grab some live audio and encode it into Vorbis, then delete the wave (because it's huge); I can cut and move frames around the Vorbis file just as well as the wave, which is all I intend to do. In executable source code "creation" in my mind applies to e.g. my Emacs undo and kill buffers, and bash history, none of which I distribute. The preferred form of modification is not always the same as the preferred form for creation. I keep the latter for modifying stuff; I don't always (probably rarely) keep the former, and I don't expect other people to either. I don't think it's fair to tell upstreams to save more state than they usually do, just to distribute the state that they do keep. I do agree that having the preferred form of modification -- the form the author uses for modification -- is absolutely essential. -- Joe Wreschnig <wres0003@umn.edu>
Attachment:
signature.asc
Description: This is a digitally signed message part