Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!zaphod.mps.ohio-state.edu!unix.cis.pitt.edu!dsinc!bagate!cbmvax!bj From: bj@cbmvax.commodore.com (Brian Jackson) Newsgroups: comp.sys.amiga Subject: Re: The Amiga featured in Computer Chronicles Message-ID: <15975@cbmvax.commodore.com> Date: 20 Nov 90 10:29:09 GMT References: <1990Nov17.150858.55@vax1.mankato.msus.edu> <15938@cbmvax.commodore.com> Reply-To: bj@cbmvax.commodore.com (Brian Jackson) Organization: Commodore, West Chester, PA Lines: 30 In article mwm@raven.relay.pa.dec.com (Mike (My Watch Has Windows) Meyer) writes: >In article <15938@cbmvax.commodore.com> bj@cbmvax.commodore.com (Brian Jackson) writes: > >(The Host wanted this to be done automaticially) > >(This is "impossible" the computer whoud have to KNOW What parts ARE wings.) > Not really. The ability to "know what is a wing and what is a beak" > requires some rudimentary edge detection and that ... > >I talked with Leo about this several years ago, basically suggesting >what you did. He wasn't sure it could be done. What is clear is that >it can't be done reliably, and TAS is for naive, inexperienced users. >What is perfectly acceptables for professionals at TBS may not be >acceptable for such users. Oh yes, I totally agree. This was just theory mind you :) We had such code working on an old Inovion PGS-1 a while back (there is a picture of this machine in the dictionary under the word "slow") that did it fairly well. As long as a given block of color overlapped at least part of the position it held in the previous frame, it could find it's way through about 95% of the time. It still stopped at each frame to be sure that you were happy with things but it worked. bj >