From the trenches

The Advisor called me today after I left her a message. I tried to answer the mouse.

On the up side: The draft should be done by suppertime.

Also, I learned that as a part-time student, I can get Adobe Acrobat 7.0 Standard for only $99 instead of $299. That really rocks, and I must at once go get a copy of my transcript before I am no longer a student.

On the down side: There will be a section of handwaving. Hopefully, by the time of the defense, I will have something to back the handwaving up with. But I don't like having to put handwaving into the draft, even, because it primes the committee to think that you're trying to pull a fast one on them.

Other fine works of research

I keep a link to this thoughtful analysis of an experiment about electron band structure in germanium by Lucas Kovar. It seems relevant, today.

The suspense continues!

So... the torque data.

Really, you don't need to know what I mean by "the torque data." It's some data, just a subset of the overall data I'm looking at. A thing.

A very uncooperative thing. It should have done X. But it didn't do X, it did Y. So I looked at it for a good long while and said, "Oh! I bet I know why it didn't do X. I need to add some Xness, and then it will do X."

So, add some Yness and sit back. And, yea verily, the data comes back a little more like X but still pretty Y. Hm.

"Well, maybe it's because I'm also asking for Z, and Z can undo X. It shouldn't have to, but maybe that's the case? Let me remove Z and see what happens."

And while that's running, I look at the code again. I've been looking at the code, with a vague feeling that there's something wrong. I check the graphs that gave birth to the code. Check. I check the math I did when I solved for A in terms of B instead of B in terms of A. Yes, all in order.

Math is tricksy. Sometimes, you take squares and suddenly all the negative numbers go away. Was it something like that? I decided to plot A in terms of B just in case. No, it looked exactly like the equation I had...

Except for the constant. I had A = f(B). The plot had A = (some number)*f(B).

All my other programs like this one? Compute (some number). Here, it had escaped me.

So. Well. Back to Square One.

The good news? The great news? If it does X now, I'm done. Done done done done done. DONE. I was right and my method works.

The bad news? It might still do Y. I have to wait and see. And it might take a while to see. And if it does Y, I'm back to the drawing board with trying to figure out why.

Quote of the Day

"130"

- The number at the bottom of the last page of the first "committee draft" of my dissertation

It's not done, really. The torque data did... well, it didn't do X or Y, it sort of went GARGLEBLAFFT all over the place. But I can wave my hands about and pretend that the results I have are really OK, which they aren't and the committee will know that, but the committee will also know that I'll be working like mad to get the finished results by the defense date. So it's all good. Ish. Good-ish.