Wednesday, June 16, 2010

One of those days

Some days I have the Midas touch. Some days I don't. And some days, well, I suppose the opposite of Midas would be the merde touch. It's French - Google it. Today wasn't quite that bad. I think I'm going to go with the petit merde touch.

It started early. I arrived at the paying job at my normal 0615 and settled in at my desk after a brief stint brewing the coffee and yelling at the newspaper. I had left a job nearly done yesterday afternoon and put the finishing touches on it this morning. It's a clever piece of code that goes into the database and retrieves an entire hierarchy of data. In other (equally opaque) words, it retrieves a top, root object, then retrieves the children of that object. As it retrieves each child, it also retrieves their children. If you were to draw out the relationships, it would look like an org chart or family tree. The benefit of getting everything all at once is that I can get at the entire family, if you will, after just one operation against the database. Efficiencies are gained by doing so.

Well, efficiencies are gained by doing so if (and only if) it actually works.  It didn't. Instead of returning my happy family of data, the program crashed with a stack overflow. What that means in a nutshell is that my computer ran out of memory. With the kind of memory computers have these days, though, that would have had to have been one very, very large family. No, it meant that something was horribly wrong in my code.

It came down to a problem with cousins, to drag the familial analogy along even further. After quite a bit of hair pulling and desk pounding, I realized what was happening. When I was creating one of the children, I had it also create its own children. As long as I was doing that, I thought I'd be clever and have it create a somewhat unrelated relative that I thought might come in handy later. You know, as a kidney donor or something. I'll call that new relative a nephew - it's nicer than "parts bank."

The problem was that the nephew was never intended to be included in the family, so as part of his creation he had to find out who his uncle was to maintain at least some semblance of association. So, when the uncle created him, he in turn tried to create an uncle for himself. That uncle tried to create another nephew, who also tried to create an uncle, who tried to create a nephew.....  On and on and on, until there's no memory left to hold it all. Stack overflow, scream of frustration, pound the desk. You know it's bad when coffee is the calming influence on your morning.

I doubt if that made much sense, so suffice it to say, it was all quite frustrating.

Once I figured out what the problem was, it was easy to fix. I neutered the uncle. Serves him right.

Unsurprisingly, my work on the airplane tonight ended in frustration. Not the screaming, desk pounding kind. More of an "Oh well, it's just been that kind of day. It will look fine tomorrow."

It was to be a light night. All I had left to do in the seemingly interminable Section 21 was to rivet on the very last bulkhead. It took four nutplates and 16 rivets, plus a detour into the second of the two autopilot brackets, but it was pretty easy.

Being as this is pretty close to the end of this section, I kind of took an audit of completeness, as it were, basically just trying to see if I had missed anything. Here's a bit of a mystery:

It's passingly odd that there are no bushings in those holes. I went back through the plans and could find no steps instructing that bushings be installed there, so there it is. A mystery.

The second to last step is the installation of a bracket that will support the fuel flow sensor for the Dynon D-180. This sensor will measure the flow of fuel (hence the diabolically clever name) through the fuel lines, allowing the Dynon to display the rate at which it is being burned by the engine. Useful, that.

The bracket goes down in the tunnel between the seats.

It's kind of hard to get at down there, which is unfortunate since it has to be match drilled to the belly skin. The long #30 bit is required.

Once drilled, the bracket had to be removed for the installation of (what else?) two nutplates. That done, I clecoed it back into place and prepared to rivet it in. Looking down into the tunnel and thinking about how hard it was going to get to get the rivet puller down there was a little depressing. Frankly, it was a battle that I did not want to fight. Suddenly it dawned on me that not only would I not need to fight that battle, it would have been foolish to do so. Who wants the wart looking pulled end of a blind rivet hanging out of their belly skin?? No one, that's who! Not even when it's airplane skin we're talking about and not their own skin! 

The right thing to do would be to install the rivets up through the belly skin. There's no easier time to choose do the right thing than those rare occasions when the right thing is also the easy thing.

Well, if not easy, at least easier.

So, last step. Looks easy: install six panhead screws up through the belly skin to act as attachment posts for some device to come later. Six screws, six metal locknuts. Nothing to it, right?

You know what they say are the three most important aspects of real estate are, right? Location, location, and location. These screws are not in good locations. Not.At.All.

After futilely struggling with the first one for twenty minutes, a thought struck me: this is the easiest one.

I uttered a sharp, non-petit "Merde!" and called it quits for the night.

No comments:

Post a Comment