#14 GGI 3.0: modesetting rework a la KGI Degas

Jon Taylor

A lot of the abstractions of modesetting in LibGGI are very well though out. IMHO, however, they don't hold a candle to the LUT-tree based approach degined by Steffen Seeger in KGI 0.9 Degas. The GGI approach usually involves bitmasks, shifts and semi-standard mode #defined types with fixed-with words, while the Degas approach involved explicit constrruction of combiner ways using LUT (lookup table) analogies. An infinite variety of word schemes representing almost any type of pixel or glyphic form could be easily constructed and mapped to a set of more traditional. It wouldn't be to hard to lift this system from Degas into GGI 3.0, and it could easily be typed or cast (as Degas also did) into a set of more traditional mode types which would be easier for the bulk of system graphics code (especially at the OS level) to use and be easily understandable to most programmers. SOmething to think about....


Log in to post a comment.