Bill wrote: > > This is a somewhat interesting topic. In my limited > > experience of having played with these a bit, I've found that > > it seems to matter what the 'context' is in which the 'operations' > > are considred as being applied -- ie. wether one is considering > > an op as 'drawing' a (possibly patterned) geometric figure, > > or as 'compositing' some color data (to a 'dst'). > > The "context" in this case is decided by which of the src, mask, > and clip have the shape you are talking about: > > src = compositing > mask = geometric figure > clip = don't change any pixels outside of this, used for windowing, > etc. That would be one way to view it yes. By 'compositing' ops one would mean a set of color-space operations, which by extension one would apply to src 'patterns', so that one can build new ones from others via color operations. There, one might would not want to be limited by some presumed semantic interpretation of the nature of 'alpha' or such, and would instead want a powerful algebra. Drawing of (closed) geometric figures with src 'patterns' is its own notion, and one would like a consistent, intuitive semantics for it. This might not be completely achievable by representing such figures by masks (and what is the semantics for that process?), but if it's a reasonable way to 'draw' such, then I'd say that one would want any semantics built up from 'compositing' to satisfy at least that: areas lying entirely in the exterior should not affect the dst, and areas lying entirely in the interior should 'compose' the pattern with the dst (for whatever given binary 'compositing' function). jose.