FINAL POINT: Historically, scientists down through time have picked up the research of previous scientists in order to further their own work, effectively taking the scientific findings of the previous era and running with it, like runners in a relay race handing off the baton to each other.
Should contemporary scientists ignore or cover up the fact that they're using the proverbial research baton of a bygone era? No, that would be very dishonest and misleading. There's nothing wrong with the fact they're doing it. That's how technology advances. It's a beautiful thing. They should cite their sources, though. Don't be a selfish pig, practice courtesy and honesty in what you do - respect the work of others when you benefit from it.
You said it
IMO the comparison to science is almost exactly right. A palette is a significant component of a finished image, though by itself it's nothing terribly remarkable. Just as there are plenty of unremarkable scientific studies, the results of which have been built on to create remarkable new advances. eg
shape preserving 2d/3d mesh warping (The results are nothing short of awesome. Implementing this in GIMP is currently being considered as a possible Google Summer of Code project)
This is not a matter of intellectual property (the entire concept is deceptive wankery, and creates the problems it is supposed to solve (eg. as in independent simultaneous creation of a near-identical work)), just a matter of common courtesy
(that said, I'd rather see a palette I made used, even without credit, rather than simply bit-rotting, never being used again)
just FYI - the likelihood of someone picking the exact same palette with N indexes is 1 in N*(256^3) i believe. (for rgb256)
so for a 16 color palette like arne's:
1 / 268,435,456
Your figures are WAAY off.
you would only distinguish that fine color differences if the palette was very large (256 entries is about the minimum palette size you could specify 256 levels of intensity as meaningful for); colour perception is relative. Except in such circumstances, 64 levels of intensity is closer the mark for colors we can distinguish somewhat readily. As the overall size of the palette drops, so does the amount of precision required to describe it.
So a 16-color palette, like Arne's, could have anything between 16*(16^3) [65536] to 16*(32^3)[524288] to 16*(64^3) [4194304, which is still 64 times smaller than the value you gave] . The exact precision would depend on the particular context the palette was being applied in. For Arne's palette, I would bet on 32 being about the right number of intensity levels.
