Path: utzoo!mnetor!uunet!husc6!mailrus!nrl-cmf!cmcl2!rutgers!ucla-cs!cit-vax!tybalt.caltech.edu!wetter
From: wetter@tybalt.caltech.edu (Pierce T. Wetter)
Newsgroups: comp.sys.mac
Subject: Re: QuickerDraw
Message-ID: <5405@cit-vax.Caltech.Edu>
Date: 10 Feb 88 21:04:27 GMT
References: <801@PT.CS.CMU.EDU>
Sender: news@cit-vax.Caltech.Edu
Reply-To: wetter@tybalt.caltech.edu.UUCP (Pierce T. Wetter)
Organization: California Institute of Technology
Lines: 28

In article <801@PT.CS.CMU.EDU> mkb@rover.ri.cmu.edu (Mike Blackwell) writes:
>The way I understand it, Hertzfeld made the remarkable discovery that when
>you're in 8-bit mode, you never have to check if you need to do bit shifts,
>since you're always byte aligned. Saves a couple of instructions per pixel,
>which could be substantial. Why the color QuickDraw implementers didn't
>figure this out, though, is beyond me - operating in 8-bit mode is common,
>and should have been special cased in the first place.
>

  More likely is that Hertzfeld realized that its silly to pack the 1,2,4, and
8-bit modes together. A faster way is to make each pixel a byte but only 1,2,4,
or 8 are significant. This way rather then toggling bits you can simply blow
away the byte which is much faster.
 The 24-bit boards have this problem as the present quickdraw is unable to 
simply set all 24bits by doing a long word write, rather is sets each in turn.
 Arrgh!

 rampant sillyness
Pierce Wetter
wetter@tybalt.caltech.edu

"When you are in it up to your ears, keep your mouth shut."

--------------------------------------------

wetter@tybalt.caltech.edu

--------------------------------------------