🚀 go-pugleaf

Craig Scott Lennox (len...@minilove.diag.stratus.com) writes:
+--
| ...stodgy old Sun (booo!! hisss!!) in another display of
| lack of humour, replaced the grok error message with a
| bland "unknown command" instead.
+--

[ curmudgeon mode ON ]

I've been enjoying this thread, but I agree with Sun here.
You have to look at from the point of view of the users (and
the suppliers) too.  Seems to me that the ``can't grok''
message is bound to cause a lot of grief for the majority of
customers who have NOT memorized the jargon file yet, and
for the suppliers' chronically overtaxed customer support
people who have to explain to customers why a common error
message contains a word not found in English dictionaries.

It's not so bad when cutesy jokes are found in source files,
because the majority of their readers WILL be hackers, but
I urge you implementors to resist the urge to cuteify error
messages.

[ curmudgeon mode OFF ]

And now a confession.  The worst error message I've ever
seen was my own damned fault.  When I was a cocky
undergraduate systems programmer, we ran a 360/44 PS shop,
and when a FORTRAN program failed, 90% of the time it would
write the same message:

  OA210I: PROG CK IRPT, OLD PSW WAS <string of 16 hex digits>

I got very tired of explaining to scores of novice FORTRAN
users that PS was too brain-dead to give the user any more
detail about what went wrong, and that 99.9% of the time it
was their fault, not hardware error.  So I changed this
message to a simple

  OA210I: YOU BLEW IT, OLD PSW WAS ...

The business of programming is frustrating enough without
arrogant, insulting messages like this one (however true it
was).  I forgot about this until one of my classmates
brought it up at my 10th college reunion.  It was a humbling
reminder of how cocky I was as a student, and how irritating
I must have been.
-- 
John Shipman/Zoological Data Processing/Socorro, NM/jo...@jupiter.nmt.edu
``Let's go outside and commiserate with nature.''  --Dave Farber