Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!samsung!zaphod.mps.ohio-state.edu!rpi!uupsi!sunic!news.funet.fi!assari.tut.fi!assari.tut.fi!n67786 From: n67786@assari.tut.fi (Nieminen Tero) Newsgroups: comp.sys.mac.misc Subject: Re: The New Macs: Greedy Compromises? Message-ID: Date: 4 Dec 90 06:41:01 GMT References: <1990Nov29.203507.25984@grape.ecs.clarkson.edu> <46966@apple.Apple.COM> <1990Dec2.084149.25494@world.std.com> Sender: n67786@assari.tut.fi (Nieminen Tero) Organization: Tampere Univ. of Technology, Finland. Lines: 27 In-Reply-To: rang@cs.wisc.edu's message of 3 Dec 90 18:36:06 GMT In article rang@cs.wisc.edu (Anton Rang) writes: > I thought that splines, as used in text operations, worked in very > small numeric ranges. Is there a reason that fixed-point math can't > be used? I'd guess that TrueType would use fixed-point math; it ought > to be faster than floating-point, too, at least until Motorola's FPUs > speed up a bit. ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ^^^^^^^^^^^^^^ Huh? Where did you get that one? Fixed point surely is faster than SANE floating point on a machine without 881/882, but 882 is almost as fast in floating point as 030 is in integer calculation and most surely faster than fixed point that's been emulated by integer calculations. You can verify the speed difference yourself with a copy of the MandelZot, if you don't believe my word for it. > Anton > > +---------------------------+------------------+-------------+ > | Anton Rang (grad student) | rang@cs.wisc.edu | UW--Madison | > +---------------------------+------------------+-------------+ > -- Tero Nieminen Tampere University of Technology n67786@cc.tut.fi Tampere, Finland, Europe -- Tero Nieminen Tampere University of Technology n67786@cc.tut.fi Tampere, Finland, Europe