22 Nov
2019
22 Nov
'19
2:08 a.m.
On Thu, Nov 21, 2019 at 2:53 PM Noel Chiappa <jnc(a)mercury.lcs.mit.edu>
wrote:
One of my favorite stories of the same vein was a masscomp story. We were
chasing a rare event (as I recall it was when we first were debugging
Multiprocessor stuff and it a lock order problem). But we could not get
the customers to tell us about what was happening, since the system
recovered quickly, but we might kill a process. We had done a few releases
and make a few changes but we could never reproduce it.
I never knew who it was but someone ??Jack Burness if I had to guess?? put
out a patch with a couple of error messages in Klingon and dumped a bunch
of information. Sure enough this was noticed, customer stopped, we got the
needed data, as they reported the error. But it was a high visibility
customer, so the president (Mr. Potatohead) got a phone call. Fossil (our
boss) made us swear it would never happen again, but he defended us to the
President. We found the bug ;-)
From: Arnold
Robbins
The Bell Labs guys in some ways were too.
And there's the famous? story about the Multics error messages in Latin,
courtesty of Bernie Greenberg. One actually appeared at a customer site
once,
whereupon hilarity ensued.