Newsgroups: comp.lang.lisp
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!sdd.hp.com!elroy.jpl.nasa.gov!decwrl!parc!masinter
From: masinter@parc.xerox.com (Larry Masinter)
Subject: Re: monitoring functions
In-Reply-To: barmar@think.com's message of 1 Apr 91 18:17:43 GMT
Message-ID: <MASINTER.91Apr1114553@airplane.parc.xerox.com>
Sender: news@parc.xerox.com
Organization: Xerox PARC, Palo Alto, CA
References: <14131@medusa.cs.purdue.edu> <1991Apr1.181743.18493@Think.COM>
Date: 1 Apr 91 11:45:53

Using "advise" or mechanisms like it to monitor performance often
doesn't work if the time-to-execute the advice exceeds the time spent
in the function being monitored.

There have been a number of clever schemes that attempt to compensate
for this "Heisenburg effect" (measurement interfering with phenomena
being measured), but most are pretty unsatisfactory.

The Interlisp "breakdown" program did this. (Introduced sometime
between 1970 and 75).

--
Larry Masinter (masinter@parc.xerox.com)
Xerox Palo Alto Research Center (PARC)
3333 Coyote Hill Road; Palo Alto, CA USA 94304
Phone: (415) 494-4365 Fax: (415) 494-4333
