In all my years of
computer engineering, I thought that nothing was impossible. Sometimes the
problem required some simple debugging, or some
creative new application of an old
idea. Most of the time it required that you twist your mind around so much that you think like a
toaster or an infrared sensor. Less often, you could solve a problem simply by brute force of will, trying every single
god-damned possible
solution until something finally worked. I thought there wasn't any computer probleme that couldn't be solved by a smart engineer.
That was when I was hired for the Everything machine.
It worked well enough, at first, with the big ticket items: flying cars, moon shuttles, FTL engines smaller than a phone book, giant donuts. Slightly more complex items were a bit harder: porn that never got boring, the best mousetrap possible, nuculear-powered submersiable wiccans. A very few amount of items were almost impossible to fabricate until some leap of genius occured, such as the non-replicating politician, the un-funny monkey, and the Philipinno midget spy film. (For some reason it just wouldn't get it, until someone removed the level 7 diagnostic module and replaced with a dual-operating level 6A.) However, everything was going according to schedule, more or less, and the machine was nearing completion.
I was working late one night, entering some of the smaller, more nonessential items into the Everything machines vocabulary, when I hit a snag. I fed it a new sample, and it spit something out that vaugley resembled it, but just barely. I checked the settings to see if they were correct for this type of item, and they were. I ran a diagnostic, it checked out, but it also said that it was reproducing the sample correctly. This didn't make any sense.
I pulled out the diagnostic modules, and self-checked them found that they were ok, reinstalled them, tried again, and got the same false return. I started to take the machine apart, trying each piece seperatly, and getting no results.
Morning came, and the next crew showed up. I showed them my problem, and they offered some suggestions, but none of them worked. Eventually, we called the project chiefs, and they weren't able to find anything wrong, except for the fact that the Everything machine wasn't quite producing everything.. The project halted for 3 weeks as the machine was disassembled and checked and re-checked and put back together. We were confident that it would work this time. I fed the machine the sample, and it spat out something...wrong. The same as it always had.
The machine was considered a failure. Oh, it had it's uses, but it pretty much became a curiousity, and eventually was forgotten about. A lot of people have their theroies as to why the machine didn't work, but I feel that some things are just intrinsic to the human soul, that no matter what you do, they just can't be translated. We feel them so much in our gut, that to intellectualize it is to ask for failure.
No matter what you do, there's just no way to teach a damn machine to make a waffle.