De-bugging history: the non-technical version
She must have been a pre-med. There was no other explanation. The behavior was routine: First, a few suspicious glances, brow furrowed and eyes narrowed. Then, the inevitable rise from the seat, slowly and casually, hoping we would take notice. And finally, when we ignored her, she asked, her voice a curious blend of plaintive exasperation and self-righteous annoyance, "Are you guys working on all the floors?"
Tricky question, that one. We could answer facetiously: "No, miss, as you can see, we are in fact only working on one floor. Unfortunately, we have not yet mastered the art of working in two places at once."
Or we could tell her the truth: "Yes. Wherever you go, we will follow. Wherever there is silence, we will bring noise. There is no escape."
But we usually settled for a sheepish apology and half-witted explanation. "It's an art project, you see. We'll move on to another floor soon." But we'll be back, the pointy-tailed devils on our shoulders were quick to add, you can be sure that we'll be back.
Spending ten days in the Sci Li does strange things to a person's mind, and when there are five or six people together the effect is compounded. During the first few days, somebody would think to remind us every few hours, "Do you remember when we thought we could do this all in one day?" Yes, we all remembered, with considerable chagrin. How quaint our initial ignorance seemed.
That had been the original plan, of course: one marathon day of hanging, crimping, testing, and running. Luckily, we learned on Wednesday that hanging and crimping took considerably longer than we had planned, and so we were able to devote Thursday to those tasks. Teams of three or four fanned out across the floors; two would go ahead, armed with wire cutters and strippers and prepare the horizontal cables. Test for left, strip and twist. Test for right, strip and don't twist. Wire insulation fell to the floor like bullet casings. The crimpers would follow, a tell-tale trail of little red barrel crimps leading through the library stacks. And, of course, the re-crimpers would bring up the rear, perhaps hours or days later, doctoring mashed crimps and reuniting separated wires.
And then the fun started.
Each floor required a control board, and each control board required delicate surgical attachment to the vertical cables. Tech House has never been known for its delicacy, but we tried. Despite the omnipresent darkness in those dark corners by the dumbwaiter shaft, despite the cold breath of hell that gusted on the thirteenth floor, despite the shortness of the wires available for crimping, despite the mad midnight-hour dashes down to the lobby when the doors were closing, despite all of that, the boards somehow found themselves connected?mostly. And when the boards were (mostly) connected, we were able to (mostly) test each floor individually.
Imagine this:
Logical ten, floor thirteen. Manning the base camp: Mad Haxor Ryan Evans and his trusty walkie-talkie. In the field: jack-of-all-crimping-soldering- elevator-riding-relay-switching-multimeter-metering-trades Neel Joshi and professional flashlight-wench Kali Wallace.
Field agent: Okay, Ryan. Give us a test func four on logical ten. Basecamp: Coming right up. Field agent: Uh, Ryan, did you write that function? Basecamp: You should have a bit test on logical ten. Field agent: We don't see anything.
<pause>
Field agent: Oh, wait. The lights weren't plugged in.
<pause>
Field agent: Well, that's weird.
<pause>
Field agent: Um, we're going to check the board.
<pause>
Field agent: Can you write all ones to this floor? Basecamp: Sure thing. Field agent: Hmmm?all of our 5 volts look like 3 volts. Basecamp: That's not good. Field agent: Right.
<pause>
Field agent: Maybe we should wait for Soren to get here. Basecamp: Sounds like a plan. Field agent: We're coming down.
And so we waited for Soren, and we waited for Brett, and we waited for news from Professor Patterson. The scent of solder filled the air. An oscilloscope blinked mysteriously. Ground wires were chopped and soldered. We asked, "Um?did we remember to turn the 12 volt power supply on?" We cursed the boards and the chips. We decided that the bus was female, because there was no other reason for the illogical flaky behavior. We came to know bits 8 and 9 on logical 10 very well. We regularly wondered, "Is anything burning?" We napped on the third floor. We had conversations through the dumbwaiter shaft. We accomplished a mind meld such that the numbers "nine" and "twelve" became interchangeable without confusion. We rode the staff elevator 12,347,065 times. We picked up conversations between children and parents on our radios. Eventually, we started to amuse rather than annoy the people who were studying while we tested. We were "those people who run around the library with walkie-talkies." The library staff knew to wait for us after the closing bell.
Despite a lot of back-pedaling and needless over-re-correction of nonexistent problems, it slowly came to life...and then died. The bits tests clicked merrily through each floor...and then didn't. Weird patterns of quasi-non-deterministic bit behavior evolved and were noted with concern.
Field agent: Hmmm?when bit 0 turns on, bits 8 and 9 turn on, but when 0 turns off, 8 and 9 flash on briefly, but only until bit 1 comes on, and then they flash off again, and bits 2 through 7 flash on, with the exception of bit 5, which doesn't do anything-is bit 5 plugged in? bit 5 wasn't plugged in-and when the test gets to 8 and 9 they go on and off correctly, but all the other bits click randomly on and off while they do that. And now bit 7 is just staying on. It doesn't even flicker anymore. Basecamp: Um... Field agent: Well, I'll go check the board.
we twisted the boards; we jiggled the wires. We didn't know why it wouldn't work. And when it worked, we didn't understand that, either.
So when it finally did work?Well, it worked. The Tetris gods stopped messing with our minds and let us play for a little while.
-- Kali Walace