Thursday, March 28, 2013

Last Minute Specification Changes

We have all been there. We have been working on something for weeks. We have it all set up, with everything functioning perfectly. And then one day, just as you are about to be done, your boss comes in and declares "Actually, can we do this other thing we hadn't planned on that has these additional requirements?" You throw up your hands and complain about how all the work you just did has to be redone... and then your boss looks at you sharply, you turn around, and start redoing it. This kind of thing happens all over the place, in IT, in business, heck even in the military ("no plan survives contact with the enemy").

You would think that science being, well, SCIENCE, this wouldn't be so much of an issue, but it happens more often then you expect. You are setting up an experiment and it occurs to someone, "Hey, we could measure this other thing too, and get a two-for-one setup". Science experiments are expensive and often overspecialized. It can cost thousands of dollars to make setups to test only one thing, often out of parts that need to be custom made and are virtually useless for anything else. That's not even including the mammoth machines needed for modern High Energy Physics, like the Large Hadron Collider (this isn't why I am not in HEP, which I'll get into in another post, but it was sure a factor). So when you are able to make a setup that can be used for multiple experiments, or at least to measure more than one thing at once, you jump at the chance.

The annoying part is when you figure this out when you are 99% done. In my case, we found a way to measure fluid velocity in our setup. Don't laugh, measuring this when you can't just shove the water through a flow meter is much harder than you think (if you are brave/mathematically savvy, have a look at this). We were going to have to do a dozen test with different samples and techniques in order to get that velocity. Now we can just use a laser. So changing things to let us do this saves us a ton of work later, and gives us a better set of data. The rub, of course, is that none of the work done to this point had a laser beam anywhere, and so it's back to the drawing board. Just like in software, it's not as simple as just tweaking a variable.

So raise a glass to me as I start over, and I will raise one to you the next time you get this dreaded visit. And remember my credo: "At least they aren't TPS reports."

No comments:

Post a Comment