Path: utzoo!mnetor!uunet!husc6!mailrus!tut.cis.ohio-state.edu!bloom-beacon!gatech!purdue!decwrl!labrea!polya!rokicki
From: rokicki@polya.STANFORD.EDU (Tomas G. Rokicki)
Newsgroups: comp.sys.amiga.tech
Subject: Re: Idea
Message-ID: <2699@polya.STANFORD.EDU>
Date: 27 Apr 88 18:25:00 GMT
References: <8804231647.AA09390@jade.berkeley.edu> <5809@well.UUCP> <9289@agate.BERKELEY.EDU>
Reply-To: rokicki@polya.UUCP (Tomas G. Rokicki)
Organization: Stanford University
Lines: 29

> Speaking of kludges regarding the date, sorry for the miniflame, but:
> both Peck (the book's author) and Rokicki (who wrote this particular piece
> of code) should be *ashamed* to have released a piece of code like
> ShowDate() on page 42 of the Programmer's Guide to the Amiga. I thought
> it was a publisher's misprint at first, it was so full of magic meaningless
> constants.

Naw, ShowDate() is actually pretty neat!  (Of course, I wrote it.)
It was *intended* (by me at least) to be a magic piece of code;
the puzzle was to figure out how it worked.  Ditto with my blitter
LIFE.  Anyone could and can write the straight-forward approach; I
wanted an approach which would use the mininum number of bytes and
execute in the quickest time.  Not for any real purpose (except to
save a couple of bytes) but simply as a puzzle.

> The algorithm is *totally* opaque, despite the fact that I understand
> the USUAL approaches to date algorithms, including the the one for
> calculating the day of the week that any day in history falls on, in your
> head. So I tried to prove it didn't work. Well, I was wrong, it does work
> (as far as I could tell). I even started to get an idea of why before I
> gave it up as a waste of time.

The algorithm in ShowDate() is very close to the one for calculating the
day of the week that any day in history falls on.  I'm sorry you had
such difficulties figuring it out; maybe I'll explain it to you at
DevCon.

Programming wouldn't be half as fun if I couldn't come up with gems like
that.