I have had the most unproductive two weeks of my life. Faced with a fairly simple problem, I’ve fairly well sat here like a total muppet, staring at a screen, achieving nothing. It wasn’t even particularly complicated. Take some code that I wrote about a year ago, that didn’t work, and fix it, so it did work. Simple enough – the code was all there, it was all running, it just wasn’t doing what it should have been. All I had to do was work through the various functions and bits of theory in there and figure out what was wrong. And there’s not even that much code, maybe 1000 lines, split up across 5 or 6 functions.
After a week, I was still staring at results that weren’t right. Little tweaks here and there had optimised my last year’s quite immature approach to C++, and so I was seeing speed increases, but speed is irrelevant when the results you’re getting are a pile of shit. Heck, I even got the thing running on one of the distributed computing services (an 8 box Linux cluster) and I was seeing improvements all round. Just no good results.
Finally, I gave in, and approached my supervisor for help. I told him I was sure the code was technically sound, and that it was the theory behind it that was wrong and needed changing. No, he said, it must be the code, try a few things here and there and let me know. I wonder if it’s because he was in part responsible for the theory?
I went away, made the changes, tried his suggestions. No joy. A week and a half after I’d started on this, and I was no closer to a solution. Determined, I started tweaking the theory. Minor changes here and there. One fairly important change here. One that was, essentially, the opposite of what he had told me to do in the first instance.
And the result? Perfect results that exactly matched our expectations. After all his insinuations about my crappy coding, and it wasn’t me at fault.
Well. It’s not about blaming anyone. But put it this way; I’m looking forward to our meeting on Monday morning when he asks me how I got on.
Leave a Reply