The journal of irreproducible results

Somewhere during my Masters, my friend and I were working on something. We had an identical environment setup on both our machines, and yet, our results were nowhere similar let alone identical. What was worse was that we couldn't even replicate our own results when we took another trial. Every trial yielded different values. We had two choices: tweak the data and report the results as they were supposed to be, or try and find an explanation that enveloped the variance in results. As much as we did try for the latter, it just didn't happen. We were working with an ambitious professor who was breathing down our neck to send in our work to a conference/journal. How do we publish when we don't even get reliable results? we asked of him in desperation. He laughingly said, If that's the case, we can always send it to the 'Journal of irreproducible results'. Although we eventually figured it out,his words remain in my head till date.

I was reminded of this because of the manifestation of a weird phenomenon on my laptop. After the whole water scenario(which you can read about here), my laptop came back with a few dents and scratches, but nothing major. Except this. I would be in the middle of something, when it would just shut down without any warning whatsoever. Thereafter, when I again turned it on, it would repeat the same with an annoying increase in frequency, till I would just give up in despair and stop working for the day. Come next day, it would boot up and work just fine.

I prowled all over laptop forums looking for something along the lines of what I was going through. I installed various packages and ran tests hoping to find some malfunctioning sensor. Alas, no such thing. Reinstalled drivers for everything that could even be remotely related. Pored through system logs till I thought of my entire life in terms of System V Signals. Worked on straight AC power without the battery. A different battery. No dice. Finally, I did what all embedded folk do---attribute it to a hardware defect :P

I got it checked by a laptop repair guy-cum-enthusiast, who monitored my laptop for three whole days, throughout which it was on its best behavior (like that annoying kid that pretends to be well-behaved amidst relatives:D) Upon my insistence, he took it apart and checked everything, indulging me as if I were an eccentric aunt. He gave it back to me, promising that it was fine.

The moment I got it back with me, it went back to its former bizarre self. I have never believed in Murphy's law as strongly as I did through that month. I would be in the middle of something that had been going on for hours, a kernel compilation for instance, when poof. Power off. Another make -j8. It was almost as if it did it on purpose. It would go off at precisely the most (in)appropriate moment, swallowing up an entire article I'd been writing without hitting save or in the middle of a demo/presentation. I was so frustrated that I began drawing analogies between this and death talking to my friends about how ephemeral everything was(how cliched right?).:D

Now, that phase of life is over and my life doesn't depend critically on my laptop. And guess what, the frequency of shutdowns has gone down drastically. Those days, it used to vary from thrice a day to twice a week. And now, it's been twice in the last two months. Murphy's law, I tell myself. But it keeps gnawing at my mind every time it happens. Why can't I figure it out? Are my debugging skills that bad? Arrrgh. I sometimes dream of maintaining a log of when it happens and then finding a pattern that would solve the mystery of life,death and everything in between. :P But until that magically happens, I am stuck with this mystery.


Looks like I might have something for the journal of irreproducible results after all, Sir!

P.S: Btw, I just found a humorous version of the journal. Here's a sample.
http://www.jir.com/pickles.html