"Michael K. Edwards" <[EMAIL PROTECTED]> writes: > On Wed, 02 Mar 2005 13:16:44 +0100, Måns Rullgård <[EMAIL PROTECTED]> wrote: > [snip] >> No, for a photograph the source is the actual physical object you've >> made a picture of, so a photograph can never be free. Either this, or >> a photograph should be considered as source. > > I really, really hope this is sarcasm, or reductio ad absurdum, or something.
Something like that, yes. >> In your case, your best bet would probably be to provide the >> photograph without the text, or (even better) provide the image in a >> more advanced format (e.g. XCF) with the photograph and text in >> different layers. > > Er, reality check? This is the software industry, not the publishing > industry. It's a pain to work around obscured data and > compression/decompression cycle artifacts when, say, fixing a spelling > error in overlaid text, but amateur image manipulators do it all the > time. If an image isn't permitted in a source tarball unless there's > a color-glossy-magazine level of professionalism in facilitating later > modifications, then you might as well toss out 98% of the GUIs in > Debian, not to mention 99.5% of closed-source software. > > It's good to encourage people to use sophisticated workflow when > creating images, as when creating software. But we don't call > software non-free when it isn't developed using Extreme Programming > methodology or UML modeling, not least because these techniques are > overkill for most module-scale programming projects. And we shouldn't > call images non-free just because they weren't shot Camera RAW, > imported to a Photoshop clone, and manipulated losslessly at every > stage. I didn't say we should be *requiring* it. I was merely stating what I consider can reasonably be called "source" for the hypothetical JPEG with overlaid text. -- Måns Rullgård [EMAIL PROTECTED]
