Path: utzoo!utgpu!watmath!att!rutgers!ucsd!usc!cs.utexas.edu!asuvax!anasaz!qip!shino
From: shino@qip.UUCP (Rei Shinozuka)
Newsgroups: comp.sys.mac
Subject: Re: Why is Excel 2.2 is a magnitude (or two) slower than Excel 1.5
Message-ID: <1415@qip.UUCP>
Date: 16 Nov 89 14:41:21 GMT
References: <7940@ditmela.oz> <4016@ur-cc.UUCP> <14211@jumbo.dec.com> <7990@ditmela.oz> <32578@ucbvax.BERKELEY.EDU>
Reply-To: shino@qip.UUCP (Rei Shinozuka)
Organization: Anasazi Inc, Phoenix AZ
Lines: 20

In article <32578@ucbvax.BERKELEY.EDU> oster@dewey.soe.berkeley.edu.UUCP (David Phillip Oster) writes:
>Note the title of this post. Guys, an order of magnitude is a factor of
>10. Two orders of magnitude is a factor of 100. A factor of two is not an
>order of magnitude (or two).

Actually, the more normal use of the term "order of magnitude" with respect
to computer algorithms is O(n).  If 2.2 runs 10 or even 100 times slower
than 1.5 for an arbitrary n, it may still run in the same order of magnitude,
with a constant of 10 or 100, respectively.  To empirically determine the
complexity of 2.2 vs 1.5, one would have to determine running time for 
a large range of n for both (and this does not even take into account data 
types and layout, machine configurations, etc)  If the running times
remain relatively proportional for both versions for all n (2.2 is 
consistently x slower than 1.5) then the programs are linear with respect to 
each other.

-rei

p.s. It's coming up on 10 years since I took my first college computer
     course, so the above may be a trifle rusty.