Re: A small stress test

• From: Richard Tobin <richard@inf.ed.ac.uk>
• Date: Fri, 30 Nov 2007 23:04:04 +0000 (GMT)
• To: "Innovimax SARL" <innovimax@gmail.com>, "Richard Tobin" <richard@inf.ed.ac.uk>
• Cc: public-xml-processing-model-wg@w3.org
• Message-Id: <20071130230404.DAABE29A1A6@macpro.inf.ed.ac.uk>

> It looks to me more an XSLT stress test, isn'it ?

Each xslt call just reduces the lowest level <a>s.  The recursion is
done by the recursive call to the pipeline in the <when> branch of the
<choose>.

In my implementation, every step is a separate unix process.  Each
level of recursion involves the pipeline itself, and the choose (which
runs the chosen branch in the same process).  So there are two
processes per level of recursion.  Ackermann's has an amazing depth of
recursion for even small arguments.

A fixed-point-finding program like this would benefit enormously from
tail-recursion optimisation, but I don't think it's a common case.

-- Richard

Received on Friday, 30 November 2007 23:03:31 UTC