Subject: Re: Lehman Report, Software patents, and more
From: La Monte H Yarroll <piggy@hilbert.maths.utas.edu.au>
Date: Fri, 26 Aug 1994 11:22:02 +1000

"L. Peter Deutsch" <ghost@aladdin.com> writes:
...
> While I'm on the subject, I would note that end-user expectations in
> the real (PC and Mac) market are much higher than in the Unix world.
> How many of the people reading this newsgroup work regularly with PC
> or Mac end-user software?  When I worked at Sun, it was just amazing
> to me how many highly experienced Unix people didn't have any
> experience with or concept of the real end-user world.  I'm not that
> much of an end-user myself -- my experience is limited to a dozen or
> so MS-DOS and MS Windows packages.

During the last two years I've been thrust full force into end-user
software on that platform which RMS refuses to develop for.

In many ways end-user expectations in the Mac world _are_ higher than
in the Unix world.  Clearly a lot of effort goes into addressing user
interface issues.  I find that I can use most Mac software without
ever looking for a manual.  I'd never think of using Un*x without
online man pages.  I've been very impressed with much of the system
administration features on the Mac.  Device driver and system
extension installation is as simple as dropping the appropriate icon
somewhere near the system folder.  This beats the stuffing out of
doing a full kernel config.

OTOH, there are two areas where expectations of Mac users are very
low:  reliability and user extensions.

It appalls me that a user application can (and often does) crash a
machine with a 68040 in it.  MMU technology is ancient.  How often do
you get a core dump running Unix applications?  If it happens it is a
most exceptional event.  I find myself having to reboot this Mac 2-3
times per week to get it out of locked states.  Other users around the
department don't seem to think this is a big problem.  Almost every PC
user I've encountered knows the <Ctl>-<Alt>-<Del> salute quite well.
It appears that lockups are considered par-for-the-course.

Mac applications are easy to learn, but then what?  If there's
something you want to do that the original programmer didn't think of,
you're stuck.  There's no shell, not even the equivalent of an rc file.
An increasing number of applications provide Preferences menus, which
are a step in the right direction, but they tend to provide only inane
options such as color preference.

Applications simply DON'T talk to each other.  The closest thing to a
pipe is the Mac Clipboard, which can only be used by hand.

I know that this extensibility problem is finally diminishing.
Scripting tools are beginning to appear.  The latest version of the
Mac OS has hooks to allow applications to send each other messages.
But my point is that users do not expect these capabilities.  They are
content with the current situation of "the computer [i.e. programmer]
knows best".

Another area where Mac and PC users' expectations are low is, strangely
enough, support.  I feel this is related to expectations about
reliability.

Over a year ago I tried to report an easy to replicate bug in an Apple
product.  A certain application would produce windows that were off
the physical screen, and hence inaccessible.  After two weeks I got a
message back from an Apple support person telling me that there were
known incompatibilities between the version of their product I was
using and the version of their OS I was using, and that I would have
to upgrade several of my packages,  "Until this is done, it would seem
pointless to try to diagnose this problem further."  After a great
deal of mucking about, I reproduced this simple bug with the requisite
upgraded software, which I reported.  I got no further response.  Two
subsequent releases of the software still exhibit this bug.

Nobody is ever surprised by this story.

This is in sharp contrast to experiences I've had with free software.
Several years ago, bash core dumped on me.  It generated an automatic
bug report; within two hours, Brian Fox emailed me a fix.

You might argue that these differences are due more to differences in
scale rather than distribution philosophy.  You might be right.

Free software generally isn't as good as commercial Mac software 
addressing user interface issues.  But a lot of free software is a lot
more reliable, extensible, and better supported than most Mac
software.