Flaming at Tom Keller

[geoff at desint.UUCP]

In article <227 at umcp-cs.UUCP> chris at umcp-cs.UUCP (Chris Torek) writes:

> Change for the better is
> a noble goal, but abuse and insults are unlikely to acheive it.
> (Please note that I am not speaking of Forrest's article, but rather
> some of those that preceded it.)

Nor, I should point out, is he speaking of Tom Keller's original posting.

I think it's about time we reviewed a few FACTS in this case.  I have seen
a lot of people calling Tom an ungrateful crybaby, as well as some
names that were much less polite.  In defending him, I have received a
few tart responses myself, though nobody has called me names.

1.  The software that prompted Tom's posting was originally offered
COMMERCIALLY for sale on the net.  The author was asking $250 for
a tape, mostly to cover his expected distribution costs.  Later, he
generously decided to forego the price and distribute via mod.sources.

2.  Both the original (commercial) announcements and the README
files that came with the software contained a statement to the
effect that it worked on one manufacturer's proprietary UNIX,
and had also been "ported to Bell and Berkeley, and should
run on both systems without any modifications" and a statement that
it should port to XENIX easily, though it hadn't tried (Tom's system
is a XENIX one).

3.  Other than these statements, there were no statements made about
any system or compiler dependencies in the software.

4.  In fact, the software is quite system-dependent in a number of
ways.  Long names were only one of the dependencies;  even after crossing
that I hurdle I had to fix what seemed to be an endless string of
null-pointer bugs before the software even came up without core dumping.
There were also a number of minor makefile problems.

Now Tom's posting was, admittedly, a cry of pain and an expression of
frustration.  Having gone through the same six-hour porting process to
try to make use of a software that sounded really neat, I can sympathize
with his frustration, although he probably should have waited a day before
posting.

But let's remember that all Tom ever asked was that system dependencies
be marked.  Those of you who have flamed back along the lines of "do you
expect us to cater to every minor system in the world?" are guilty of
incorrect reading and should return to primary school for remedial
instruction.

And let's also remember that Tom wasted a large amount of his own time
trying to bring this software up, and wound up with nothing to show for
it.  A little better "product labeling" would have allowed Tom to
realize that his effort would be wasted and spend it somewhere else.

Those of you who have said "if you don't like it, ignore it" are
ignoring this nontrivial issue.  I don't have a lot of time to waste.
I don't really think it's asking very much to request that people give
me enough information to make an estimate of how much time a particular
program is going to cost me, so that I can make a cost/benefit tradeoff.
(I have tens of programs from net.sources that I have never unpacked
because the cost of un-sharing them is only reasonable if I have need
of their particular functionality).

If the author of this software package had been a little more
sophisticated, he would have realized that there are many systems out
there that still suffer from seven-character names, as well as knowing
about NULL pointers.  It's too bad that Tom posted before he lost his
frustration, but let's not lose sight of the basic request:  *please*
mark all known system dependencies in your software.  If you don't know
what they are because you are fairly new to the UNIX world, a good
substitute is to list the EXACT hardware/software configurations the
software is known to work on, so that those of us who are more experienced
can make an intelligent guess.

Oh, and one more thing.  LONG before I ever defended Tom, I got the
software up and running on my system.  The next thing I did was to mail
documented context diff's to the author so that his next posting would
be more portable.  So get off my back, okay?

If anybody wants diff's to solve long names and null pointers in the
software that frustrated Tom, let me know and I'll mail a copy.
    
-- 

	Geoff Kuenning
	{hplabs,ihnp4}!trwrb!desint!geoff