Path: csiph.com!v102.xanadu-bbs.net!xanadu-bbs.net!feeder.erje.net!eu.feeder.erje.net!eternal-september.org!feeder.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: Paul Rubin <no.email@nospam.invalid>
Newsgroups: comp.lang.forth
Subject: Re: WITHIN
Date: Sun, 09 Mar 2014 01:24:22 -0800
Organization: Nightsong/Fort GNOX
Lines: 33
Message-ID: <7xob1fbu7t.fsf@ruckus.brouhaha.com>
References: <52e54d17.461102185@news.demon.co.uk> <7xha7hiv3w.fsf@ruckus.brouhaha.com> <CLGdnfOUcLjxyY_OnZ2dnUVZ_oCdnZ2d@supernews.com> <7xr46lruq1.fsf@ruckus.brouhaha.com> <o8ednawHAOPUk47OnZ2dnUVZ_tGdnZ2d@supernews.com> <7x7g8cvagd.fsf@ruckus.brouhaha.com> <KYmdnW2ZKqjp2InOnZ2dnUVZ_tudnZ2d@supernews.com> <2014Mar6.190908@mips.complang.tuwien.ac.at> <TYqdnSjeJcmsmYTOnZ2dnUVZ_oydnZ2d@supernews.com> <2014Mar7.135132@mips.complang.tuwien.ac.at> <7x7g85him2.fsf@ruckus.brouhaha.com> <2014Mar8.170211@mips.complang.tuwien.ac.at>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Injection-Info: mx05.eternal-september.org; posting-host="d94d289a4df6ae47ea4d4f8b2ae808e7"; logging-data="28408"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1/2UINqatHZFxTNFX306BwU"
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.1 (gnu/linux)
Cancel-Lock: sha1:WTFSrbaYvu+reLmWmsDau2/9gBg= sha1:LIWyke8q8VLkI+D/DZQlLakw8Uc=
Xref: csiph.com comp.lang.forth:28992

anton@mips.complang.tuwien.ac.at (Anton Ertl) writes:
>>    : endpoints { n a b -- a<=n n<b } a n <=   n b < ;
>>    : outside ( n a b -- flag ) endpoints or ;   \ precondition: a > b
>>    : inside ( n a b -- flag ) endpoints and ;   \ precondition: a <= b
>
> Looks overfactored to me.  In particular, I don't see ENDPOINTS as
> representing a concept.

Hmm, I thought the idea was to avoid duplicating code.  I guess I could
have called it CHECKBOUNDS, since it compares the test point against
the two boundary points A and B.

> In any case, I think that the proper WITHIN is superior in every way.

I'd agree it's a better implementation for 2's complement integers.  It
seems to me there are plausible use cases for floats, strings, bignums,
and various other types.  There was even a Forth-like cpu that used
ternary instead of binary for integers:

  http://brokestream.com/daf.txt
  http://en.wikipedia.org/wiki/Setun

> If you have trouble following it, see it as an incentive to learn
> about modulo arithmetics.  

In my case it made sense after it became clear (from Andrew's post) what
the function was actually supposed to do.  It was more obscure in the
ANS glossary, in part because of the stack juggling in the reference
implementation, and in part because the verbosity of the documentation
made it sound more complicated than it was.  The ANS glossary might not
be a place for it, but in a tutorial I think it would be clearer
exposition to present the generic version first, then the optimized
version.