[geeks] my head just went explodey

[Charles Shannon Hendrix]

Sat, 02 Apr 2005 @ 22:52 -0800, Gregory Leblanc said:

> It's exactly this sort of crap that drove me -out- of the industry.  All
> the micro-management by folks who don't know Jack Schitt, and hiring
> filters on only the right buzzwords.  At the moment, I'm in school
> getting an engineering degree, in an engineering discipline.  I have
> hopes that folks are more clued here that in the computer fields.

My first job out of college was doing realtime work, with a lot of
electronic engineering that I had to learn (or try to) the hard way.

In some ways, it was better because the laws of physics and the
requirements for timing dictated a lot of what we did.

However, don't for a moment think that PHBs can't screw that up.

I hope you never meet an engineering PHB, but be warned, they are out
there.

In 1994-1995 I worked for an engineering firm on civilian coal and
nuclear power plants, and US Navy submarine power plants.  It was fun
work, but a PHB from hell made every day a chore.

For example: 

I wrote a C routine to interface to a FORTRAN shared memory database.
The C routines ran on VMEbus '040s and VxWorks, the FORTRAN ran on Sun
workstations.  The database was stored on a hard drive on the sun.
Sharing happened over ScramNet, a fiber-optic shared memory system.

My boss told me that the times I reported for my db routines were not
possible.  I was reading 14K entries in few seconds, or streaming them
in sub-second times.  

His reasoning:

    14K * 9ms (average disk access time) = 126 seconds

It gets better:

ScramNet lasers had fragile lenses, and would break if you just had a
mean thought.  When we finally got a set that wasn't so fragile, and
some shock-restraint fiber, the network stopped working after about 30
minutes.  This crap was happening during a 70 hour per week deathmarch.

Lynn (an EE I worked with) and I were so tired we were nearly delerious.
Debugging from board level up through to the software showed nothing
wrong.

Fiddling around, I unplugged a node from our fiber ring, and a bit later
plugged it back in.  Lynn said, "Hey, I see 0xDEADBEEF and 0xBABEFACE on
the debug panel, it's working!"  That was our test pattern, since the
rest looked like gibberish. 

I had absolutely no idea why that made it work, and still don't.  But I
replied, "Yeah, I noticed there was too much photonic pressure in the
lines, so I let some of the light out."

Lynn was so tired he just said, "OK, that's cool."  Then a bit later a
big grin appears on his face as he realizes what I had said, and we both
had a good laugh about it.

What happened next though, still amazes me even today after many more
years of PHB encounters.

Lynn then went down the hall, and I followed.  When I caught up, he was
in the conference room telling our boss that we'd fixed the problem by
reducing light pressure.  Not wanting to appear ignorant, boss says "I
thought it might be something like that.  Sometimes we figure it is
better for you younger fellows to figure things out on your own."







-- 
shannon "AT" widomaker.com -- ["If you tell the truth, you don't have to
remember anything" -- Mark Twain]