Path: utzoo!news-server.csri.toronto.edu!rutgers!sun-barr!cs.utexas.edu!usc!wuarchive!udel!haven!ni.umd.edu!uc780.umd.edu!cs450a03
From: cs450a03@uc780.umd.edu
Newsgroups: comp.lang.misc
Subject: RE: Runtime Polymorphism -- To Have and Have Not
Message-ID: <12MAR91.18285477@uc780.umd.edu>
Date: 12 Mar 91 18:28:54 GMT
References: <607@optima.cs.arizona.edu>
Sender: usenet@ni.umd.edu (USENET News System)
Organization: The University of Maryland University College
Lines: 53
Nntp-Posting-Host: uc780.umd.edu

When I first hit this subject chain a couple days ago, I decided that David
Gudeman was a breath of fresh air.  I wanted to follow up immediately, but
promised myself to wait till I'd read through to the end of the newsgroup
before I posted.  So, I'm responding to a large number of articles, and
hopefully I'm remembering everything right.

First, I get paid to maintain/develop code on a large system (gigabytes
of data, millions of lines of code) on what Mr Gudeman would probably
consider a strongly typed (runtime type checking) system.

Quick summary of personal experience:

The runtime errors I've run into include
          divide by zero
          array reference out of bounds
          file (or file system) full
          access denied (permissions set wrong)
          out of memory
          undefined variable

Oddly enough, these errors tend to show up much more frequently in
poor code than in good code, and in good code they generally only
show up shortly after the code goes into testing.

Oh yes, forgot a couple:
          race conditions
          data corruption

An example of the last was when I changed a file naming convention
from using a date encoded in the file name to using date as an
independent property, ran a routine to convert existing files but
had not actually installed the function that dealt with the new
convention.

I fail to see how a statically typed language would save me from
any of these circumstances.  At best, the language could forbid me
from doing whatever operation caused the problem (array index out
of bounds?  well, you shouldn't be using array indices... use our
Ronco brand array substitutes instead!)

The more general problem of incorrect logic has only one good solution:
(1) figure out what the problem is you are trying to solve
(2) figure out how to solve it
(3) implement it
(4) check the implementation to be sure it is correct

And if step 3 only takes me a few hours instead of a few days or a
few weeks, well, maybe I'm "undisciplined", but we're beating our
competition.

Tough luck, huh?

Raul Rockwell