On Tue, 29 Mar 1994 12:40:25 -0600 Virginia Metze said:
>I for one am unimpressed by claims that software is easy to modify
>and fix.  I don't expect to have to modify or fix software.
>Professionally done software does not require an in-house tinkerer/
>hacker.
Virgina:
 
You don't make a living maintaining systems software, do you?  For those of
us who do, "ease of modification and fixing" is *very* important.  I have
*never* seen any piece of significant software that was bug-free.  And I've
been in this business for 15 years.
 
Listserv has bugs.  Listserv has places where it needs to be modified to fit
into our local environment.  Fortunately, Eric designed it so that I can
make the modifications, and Eric can ship fixes that "drop in" easily.  So
if there's a problem, I can have it *fixed* within 24 hours, and have on
occasion (Thanks Eric for dealing with some of the screwy things we've broken
Listserv with).
 
As an alternate example, the IBM 'TCP/IP for VM' product is generally admitted
by all, including IBM, to have a terribly broken software service scheme.  The
result is that a site can buy a new 3172 Ethernet controller for their IBM
mainframe, and then have to invest several man *MONTHS* in order to straighten
out all the pre-requisite patches and fixes in order to install the patch that
provides support for the 3172.
 
Those interested in this topic should read Lauren Ruth Wiener's "Digital
Woes - Why We Should Not Depend on Software".
 
 
                                  Valdis Kletnieks
                                  Computer Systems Engineer
                                  Virginia Polytechnic Institute