After two days of gathering test data, the results are on a sheet before me. It's time to look at the problem in situ. I fetch some CDs and my mug, and open the log book to a fresh page. Looks like this one will take a while. Everything points to a problem in the core.

Watching the frames peel by, I ponder the changes in this line of work, and count my blessings. My predecessors might spend weeks getting to this point, with the limited technology of the time. It's still tedious, tricky and nerve-wracking, though.

25... 30... 35... it's getting very murky, and sense of direction is easily confused. There are landmarks here, too, if you know what to look for. I make careful notes as my starting point disappears from view. My watch ticks over another day as the descent slows, nearing my first candidate for the source of the trouble.

I'm not too familiar with this part of the huge, sprawling structure supporting the facilities on the surface, but no one really knows it all. Over 6 million sections, most of them custom-built, modified again and again over four years of construction.

The sun is rising somewhere as I close in on the target area, at the bottom of Hell, looking for the one piece in thousands that has finally folded up under the changing stresses from above. It may set and rise again many more times before I find and repair the damage, and return to the world of air and light.

:;Smarasderagd:;

"Specifications, design, and coding can be done at any speed -- only debugging takes time."
-- Gerald M. Weinberg.