I reiterate my belief that what is good for graphics programmers is good for
graphics users. Why? Several reasons.
(1) Graphics programmers ARE ultimately graphics users. A programmer who
isn't willing to use the finished product in this area has less vested
interest in producing good product.
(2) A good programming interface makes programming graphics easier, and
hence is likely to produce more programs, of which some percentage will be
coded well, useful and hence worth using. If the percentage of programs
which are good out of the lot remains approximately fixed, more programs
means more application programs which are worth using. I see no positive
correlation between a good programming interface and bad coding technique or
lack of usefulness of the finished product, and anyone who believes they do
is free to comment.
(3) An efficient, reliable implementation of graphics at the operating
system level means graphics work well. This is good for programmers in that
it means application problem reports are more likely to be real application
problems, and results are more likely to be reproducible. It's an incentive
to program for, and use, the particular implementation.
CAVEAT: Where I'm referring to operating system, programming interface,
etc., I'm not referring to any particular locus of implementation, kernel,
library, a combination, etc. Just ``operating system'' level as opposed to
``application level.''
I could go on, I imagine, but I think I've made my point. This is pretty
basic stuff, so consider it was intended for the logically impaired...
;)
lilo