I'm trying to determine OSS compliance for a body of source code, part of which includes sections from OpenEmbedded, and contains a lot of patch files, as they (seem to) try to adapt various tools.
In these patch files I find a range of copyright possibilities
- the patch file removes a hunk of code from an original, hence 90% of the patch file itself is the original code, so the patch file is "derived" and the original copyright still applies
- In betweenies
- the patch file contains just a few lines of actual patching (or is primarily additions), and has a large number of lines as comment explaining why it is needed. Seems fair to assign the copyright in that file to the patch author, not the original
How should copyright be identified for the inbetweenies? What lines can be drawn?
The tool I am using is insisting on attribution per file, so I have been assessing a "ratio of contribution" for original/patch authors and picking one. Is that a reasonable/fair/legal approach?
Also: I have been applying the merger doctrine the context lines in a patch file (those lines around the actual change). Hence excluding them from "what ratio is original/patch?") and that tends to bias the ratio toward the patch author, not original author - is that fair?