Path: csiph.com!x330-a1.tempe.blueboxinc.net!aioe.org!news.swapon.de!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail
From: Paul Rudin <paul.nospam@rudin.co.uk>
Newsgroups: comp.lang.python
Subject: Re: Fibonacci series recursion error
Date: Sat, 30 Apr 2011 12:29:00 +0100
Lines: 19
Message-ID: <87ei4k0ygz.fsf@rudin.co.uk>
References: <a78a0ffd-03cc-4540-addb-b8e206d165d5@n2g2000prj.googlegroups.com> <KtMup.29280$8U5.7734@newsfe20.iad> <87iptw1egh.fsf@rudin.co.uk> <up6t88-4p7.ln1@svn.schaathun.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Trace: individual.net RZXJB2D5VgmnWPY0iRWfEg5+c0ATNvvDTsFHJJFDSIDGQpn3fL
X-Orig-Path: hal9000.rudin.co.uk!not-for-mail
Cancel-Lock: sha1:nswf0WnPonN/hZbZiTq+fvM24ao= sha1:es9WZyzHTMrJ9nx7QW1m06dAlsw=
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.1 (gnu/linux)
Xref: x330-a1.tempe.blueboxinc.net comp.lang.python:4196

Hans Georg Schaathun <hg@schaathun.net> writes:

> On Sat, 30 Apr 2011 06:43:42 +0100, Paul Rudin
>   <paul.nospam@rudin.co.uk> wrote:
> :  Writing recurive code is acceptable and is a nice clear way of
> :  expressing things when you have naturally recursive data structures, and
> :  can lead to perfectly good compiled code. The problem in CPython is the
> :  lack of tail optimization, so it's not a good idea for python . Some
> :  language standards guarantee tail optimization...
>
> Well, if you run into a case where tail optimisation really makes a
> difference, you probably want to reimplement it for a compiler that
> guarantees tail optimisation, rather than reimplement it without
> recursion within python.

Clearly it makes a difference in any case where you'd hit the recursion
limit. It's no big deal to do your own unwinding of the recursion to a
loop, but the code is typically less clear.