reminds me of a story from one of my favorite books, a truly geeky must read in my opinion, "the soul of a new machine" by tracy kidder. they're bringing a new minicomputer to life, its going to be a shiny new line of 32 bit supermini's (the book has been out for awhile heh), and they're deep into the debug on a prototype, having random nondeterministic errors that are often the case with complex systems running with a large number of internal states, etc, and they're trying to recreate this random error or at least encourage it to happen more often, and the VP of engineering comes in, gets a run down of what is happening, logic probes everywhere detailing intently internal signals etc etc
The guy steps up to the prototype with the board on an extender, grabs it, and shakes it quite violently and it fails. the engineers are then tasked to rewire the thing as connections just got too loose heh
Too funny... we had something similar happen on board the submarine I was on. A relatively new piece of equipment started to suffer from sporadic errors after just a couple of months being underway. They'd trace the problem to a board and replace it only to have a different problem appear after even just a couple of minutes later.
They ran out of replacement boards and, since we were underway, there was no where to go to get new ones.
I took a look at some of the old "failed" boards under a bright light and noticed the edge connectors appeared to have a bit of a coating on them. The current failure required a board that they ran out of replacements for, so I took a good old fashioned pink eraser and polished up the edge connectors, removing the apparent coating that had built up in the process. I installed the board and fired the system up and it worked at least for a couple of minutes.
I powered down the machine, took out all the boards, did the pink eraser thing to all the boards, put all the boards back in, powered it up, and no more problems.
It was the tiny bit of amine (a chemical used to remove CO2 from the air) that gets into the air and seemed to love plating out on gold edge connectors in machines. The pink eraser was a bit too abrasive for continued use but we figured out that we could wipe the connectors with a bit of white "gillie" (our pet name for pure grain alcohol, which was used to clean compensatory plates on our sonar systems) to remove the coating and added that process to our monthly maintenance on the machine. We even built a special tool (basically, something that looked a bit like a gutter with two fine grain sponges held close together that also acted as wicks for the alcohol) to run the edge connectors through that did the cleaning "auto-magically" and quickly.
is pronounced "ree-bar
" and is a "Modenism
" for R
First step towards the paradigm shift of writing Set Based code:
________Stop thinking about what you want to do to a ROW... think, instead, of what you want to do to a COLUMN.
"Change is inevitable... change for the better is not".
"If "pre-optimization" is the root of all evil, then what does the resulting no optimization lead to?"
How to post code problems
How to Post Performance Problems
Create a Tally Function (fnTally)