Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/5/84; site terak.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!bellcore!decvax!genrad!panda!talcott!harvard!seismo!hao!noao!terak!doug
From: doug@terak.UUCP (Doug Pardee)
Newsgroups: net.works
Subject: Re: Profiling on Unix
Message-ID: <499@terak.UUCP>
Date: Thu, 11-Apr-85 12:38:52 EST
Article-I.D.: terak.499
Posted: Thu Apr 11 12:38:52 1985
Date-Received: Sun, 14-Apr-85 06:36:15 EST
References: <483@terak.UUCP> <2251@wateng.UUCP> <492@terak.UUCP> <3278@utah-cs.UUCP>
Organization: Terak Corporation, Scottsdale, AZ, USA
Lines: 20

> Gee, why don't you look at 'cc -p' and prof?

Touche'.  I'm mortally wounded.

An explanation of how I came to make such a blunder...

Nine years of my career was spent with a software house, working on
optimizing compiled code and on profiling packages.  When I was
introduced to Unix(tm), my reaction to "prof" and company was "Yecch!".
I was incredulous that in this enlightened age that a profiling package
required recompilation of the program to be profiled, and even then only
reported the results in "buckets" of PC ranges.  That kind of
foolishness is, in the profiling biz, pre-1960 technology.

But this is no excuse for my rash claim that the designers of Unix
apparently didn't think much of the concept of profiling.  I know as
well as you do that half-baked utilities are the *norm* in Unix and in
no way indicate any scorn on the part of the designers of Unix.
-- 
Doug Pardee -- Terak Corp. -- !{hao,ihnp4,decvax}!noao!terak!doug