Saturday, October 29, 2005

When the rain dance goes wrong

It's snowing. Big, fat flakes. I'm not even going to dwell on the fact that it's not even November yet. What's more screwed up is the forecast for the rest of the day/night:

This Afternoon: Rain showers likely before 3pm, then periods of snow showers. High near 40. North wind around 17 mph. Chance of precipitation is 80%.

Tonight: Snow showers likely before 9pm, then rain and snow showers likely between 9pm and midnight, then a chance of rain showers after midnight. Cloudy, then gradually becoming partly cloudy, with a low around 35. Northwest wind around 18 mph. Chance of precipitation is 70%.

... and then for Sunday:

Sunday: Partly sunny, then gradually becoming sunny, with a high near 63. West wind between 16 and 20 mph.

I think it's fair to say the weather gods are confused. So, anybody out there doing a rain dance, or a sun dance or a weather dance of any sort: Stop. You're doing it wrong and sending mixed signals. Or your weather god took one too many hits off his/her celestial crack pipe.


Tuesday, October 25, 2005

More gems from statistics class

My stats professor continues to generate priceless one-liners. Today's entries:

An unintended pun, after deciding he was tired of writing down the full formula for something, which involved a random variable:
"I don't really want to continue writing this random stuff"

And the best one yet, after a proof that was utterly opaque:
"I know it's impossible to comprehend what just happened ..."

That one made me laugh out loud, which earned me a few dirty looks from people who must have thought that I was laughing at whoever didn't understand the proof; little did they know that I was totally lost myself and had no clue what I'd just scribbled down. [For the statistically inclined among you, the proof concerned the "uniformly most powerful test for one-side hypothesis testing". I don't even fully understand what that phrase means, let alone the math behind it ;-)].

Monday, October 24, 2005

Funniest 3.5 minutes I've had in a while

Here. Part of what cracks me up is the guy in the background just working on his computer like there's nothing going on behind him.

"Damn, that was $#%@#$#@ hard !"

That should be the subtitle of the paper "Combining two genomes in one cell: Stable cloning of the Synechocystis PCC6803 genome in the Bacillus subtilus 168 genome". The paper reports on the efforts of Japanese scientists to staple together the genomes of two entirely different bacteria and is basically the scientific version of "Pig and Elephant DNA Just Won't Splice" [see here for backstory]: the vast majority of the paper is basically a report of all the stuff they tried and how it failed until it finally worked, 8 years after the project was started. That's quite a contrast to the usual biology paper, in which the "Methods" section [aka "implementation details"], is only a fairly small bit of the paper, and often banished to the very end of the paper, in small font.

One thing that surprises me a bit is that the paper doesn't really get into how the organism with the glued-together genome behaves, other than noting that it only grows in the growth medium needed by one of the bacteria. You'd think that if you created something that chimeric, you'd at least prod it in a few ways to see how it behaves, but there wasn't any of that. Maybe these poor folks just wanted to get this paper out so they could take a breather before actually thinking about what they created.

All that said, though, the effort is an impressive tour-de-force of molecular biology and another example of large-scale genome manipulation [similar to earlier efforts I've mentioned]. We're definitely getting much more ambitious with the length of DNA chunks we're trying to manipulate; once we get truly low-cost, large-scale DNA synthesis, we're going to be in a position to try some really cool stuff, and get results much more quickly than we currently can. It may be a bit like the switch from writing programs on punch cards, submitting your stack of cards to the High Priests Of The Machine and then waiting a day to get your results [which is, I'm told, how it used to be ...] to an interactive programming environment -- when the cost of trying something is lower, you try more stuff and make faster progress.

Sunday, October 23, 2005

Conferring scientifically

Over the last few days, I spent some time at the Sixth International Conference on Systems Biology. Impressions, in no particular order:

Scientists fall asleep during lectures too: talks at scientific conferences like these are no different than other talks and lectures -- there were lots of sleeping people, aided by the dimmed lights, the information overload and the fact that most scientists aren't exactly super-charged motivational speakers.

High-density-poster-induced trauma:
There were 300+ posters on display. As if that wasn't enough, some people were apparently under the impression that the right thing to do was cram all the content that's normally in a 5-10 page scientific paper onto their poster, resulting in a 2ft x 3ft [or however large these things are] sheet of paper packed with text in 10-point font, lots of diagrams with multi-coloured data and a bunch of equations. My usual interaction with a poster was to read the title, think "oh, that looks interesting", start to read the text, realize that it'd take me 30 minutes to get through it all, look around the poster for big-picture explanatory diagrams/text, not find any and then stare at the poster for another 30 secs before just giving up and moving on. I think it's fair to say that I got pretty much nothing out of that bit of the conference, mostly because I just wasn't willing to put in the necessary effort.

Mixed feelings: Overall, I found the conference both exciting and depressing. It was depressing in the sense that it served to highlight all the interesting stuff that I'm not working on, and probably will never get to work on. By the same token, it was exciting because it reinforced the fact that there's no shortage of interesting biological questions [that computational methods can help to explore]. Granted, a lot of these questions will not be explored in industry [further reinforcing my "you're not going to get to do anything in that area" bit of depression], but there's still such a huge pile of open questions in the areas important to industry that that's a minor concern.

This was my first scientific conference, so I guess it's another discrete step in my current transition between fields, albeit a rather small step.

Wednesday, October 19, 2005

The danger of exams that test actual understanding

So you're going along and feeling reasonably hoopy, even maybe like a frood. You're dimly aware that, out there, there's a danger that's doing the thing that makes danger so unpleasant: it's getting closer. [A phrase shamelessly stolen, and poorly translated, from "So zaertlich war Suleyken": "Die Gefahr tat, was sie so unangenehm macht: sie naeherte sich".] And then, one day, without much fanfare, it arrives ... and afterwards you feel distinctly unhoopy. You may even be bleeding a bit.

What I'm referring to here is the midterm in my machine learning class, an hour and a half of fear, pain and misery [for me, at least] that transpired today. I'm guessing that I'm looking at a score around 40%. If the TA's are feeling generous.

The problem wasn't that the exam was insanely difficult, it was that it required you to have really internalized and understood the material we covered. In my case, it's pretty clear that I'm sorely lacking in that department. Blergh.

[PS: They found my check. No smashing. A good thing for all concerned.]

Thursday, October 13, 2005

Bank make Hulk angry ! Hulk smash bank !

[Rant, best to ignore]
After growing increasingly certain of this in the last few months, I now have final confirmation: the branches of Bank of America that are in Massachusetts are piles of steaming dog poo. Yes, I know they used to be Fleet Bank, and they aren't quite doing merging with BoA, and merging the IT systems of two different banks is really, really hard [*sniff*] and whatever other excuses people have for doing a shitty job, but this time they've really outdone themselves.

Ever since we moved here, which was shortly before BoA and Fleet Bank merged, I've had the devil of a time depositing checks into my BoA account. The usual interaction goes something like this:

- I walk up to the teller window, give them a deposit slip and my checks
- The teller types in my account number, looks confused, retypes my account number, looks even more confused, hits a couple more keys and finally asks: "Are you sure this is your account number ? I can't find any account with that number."
- I reassure him that, yes, I actually do know my account number, and have known it for the last 8 years
- He tries typing it in again, still fails to find it, and then asks me for things like my Social Security number, mother's maiden name, blood and urine samples etc, all for naught.
- Finally, he asks "What state was this account opened in ?" [despite the fact that it's clearly marked on my deposit slip], I say "Washington state" and we engage in a ritual "Washington state .. is that DC ?";"No, it's not. It's Washington state" exchange. Apparently, most BoA tellers here didn't get the memo that Washington DC is on a different coast than Washington state ...
- Armed with every bit of personal information, he then tries once again to find my account. And fails. Miserably.
- Finally, it's time to Call Somebody Who Knows What The #$%#$ Is Going On -- the teller calls whatever geniuses they have sitting in the back of the bank, describes the situation, gets some instructions and ... sometimes they find my account, sometimes they don't. If they don't, well, then they have to call a different set of back-office oracles [or just plain back-orifices ...]. Repeat until success.

Sometimes, the process described here takes upwards of 30 minutes, at the end of which I'm ready to squeeze through the little hole in the partition and commit mass murder by shoving the little nametags that proudly proclaim "Higher Standards" down the throat of all the bank employees. The only thing that prevents me from doing so is that I'm too big to fit through the hole. That, and the fact that, deep down, I know it's not really the teller's fault that he's been given a shitty system to work with.

So, that's how it went until last week, when I made the mistake of not actually waiting around to see that my check found its proper home -- when I walked in, there was a long line for the tellers and they had somebody just taking checks off people, writing down the account numbers and then handing out receipts, to speed up the process. I handed the lady my check [despite some misgivings] and walked back out; thankfully, I had the sense to keep the receipt.

Now, a week and a half later, that check still isn't in my account. And nobody has any clue where it is. I've talked to the BoA customer service folks, to the people at the branch where I deposited the check, waved my receipt into lots of people's faces and ... nothing beyond "We're still looking into it. We'll give you a call when we figure out what's going on.". Not exactly what you want to hear when said check is a month's paycheck [which, despite, or precisely because it's a graduate student stipend, I really do need in my account ...].

A couple more days of this crap and I fully expect that I'll finally lose my already precarious grip on civility. And they won't like me when I'm angry -- it doesn't happen very often, so when it does, a lot of stored-up potential is released.

Tuesday, October 11, 2005

Motivating statistics, and vanity

[Random book review]
I just finished reading "Freakonomics", which talks about using statistics/data mining to answer various interesting real-life questions like why [apparently] lots of crack dealers still live with their mothers, how to catch teachers cheating on behalf of their students on standardized tests etc. From that perspective, I think the book could serve as a useful motivating tool for statistics professors: have their students read this book and then redo some of the analyses described, using techniques taught in class.

That said, there was one thing about the book that rubbed me the wrong way: every chapter is preceded by a breathless passage [like this] from a 2003 NYT Magazine article
about how brilliant Steven Levitt [one of the authors of the book] is. As if that isn't obnoxious enough, it gets worse -- that article was written by Stephen Dubner, the second author of the book. I'm tempted to ask whether they dislocated their shoulders slapping themselves on the back. And what the hell is a "rogue economist" [as Steven Levitt is described on the book cover] anyway ? I have visions of a skinny man in a tweed jacket and glasses snorting and pawing at the ground before charging at a statue of Adam Smith and trying to bite it ...

Tuesday, October 04, 2005

The apotheosis of nerdliness

Another entry in the "Probably only at MIT" category: I saw somebody today working out on a Stairmaster ... while using his laptop, propped on the Stairmaster's "dashboard". Maybe he was inspired by the surfboard laptop. Or he really needed to get some work done, but also wanted to exercise. Or he just thought, in that "independent-thinker"-bucking-social-conventions-way that is endemic here: "Why not ? It's really not that different from reading a magazine or listening to an iPod, it's just a slightly different form factor".

I'm waiting to see somebody soldering together a circuit, or plugging ICs into a breadboard while exercising ...

Superman reborn

3 seemingly-disjointed facts:

#1. Superman died last year
#2.
Nicholas Cage is a big Superman fan
#3. Nicholas Cage just had a son

Put these together and you get a baby called, why,
Kal-El, of course.

I guess someday the kid will be happy that his father wasn't a fan of J'Onn J'Onzz.

Monday, October 03, 2005

In the spirit of "Go Fug Yourself" ...

Spotted, on a woman in her forties [ie old enough to know better], a necklace whose construction must have consisted of a set of steps like these:

- Find a small animal, say, a terrier
- Kill it, taking care not to damage the ribcage or spinal cord
- Extract the ribcage, sternum and spinal cord, without separating them
- Chop away the sternum
- Cut the spinal cord into pieces such that each section has a rib on each side of it and looks a bit like a boomerang
- File the ends of the ribs to sharp points
- Anodize each piece to a nice shiny black sheen
- Thread pieces on a chain
- Congratulations, you've made an awesome necklace !

I mean, damn. That thing looked like a piece of Klingon battle jewelry.

PS: If you're not familiar with "Go Fug Yourself", take a look here. It actually got a write-up in the WSJ recently -- they must have been hard up for material =)

Saturday, October 01, 2005

Teaching as a [non-]career

From an NYT article about the increasing popularity of the "Teach for America" program [emphasis mine]:

---
Lucas E. Nikkel, a Dartmouth graduate, wants to be a doctor, but for now he's teaching eighth-grade chemistry at a middle school in North Carolina, one of nearly 2,200 new members of Teach for America.

"I'm looking at medical school, and everybody says taking time off first is a good idea," he said. "I think I'm like a lot of people who know they want to do something meaningful before they start their careers."
---

So:

1) Teaching is apparently not a real career.
2) Once you start your "real" career, you're by definition not doing anything meaningful.

Given the high turnover rate in teaching [at least below the university level], I suppose you could make the claim that teaching doesn't end up being a long-term career for most people who start out as teachers. And I can't say I really blame them, if Christina's experience as a middle school teacher is anything to go by -- given the conditions in most schools, you have to really, really love teaching [or not have any other options, I suppose] to stick with it.

I do wonder how these young sort-of idealists are perceived by people who are career teachers, though [I say "sort-of" because the article implies that a lot of them do this as a way to help them get to where they really want to be ie law school, medical school etc, not because they're all that excited about teaching itself]. My guess is that the Teach For America folks are regarded as dilettantes, listened to indulgently while they talk about all the amazing innovations they'll bring to their teaching and then patted on the head and told to run along and play.

On the flip side, while they may not have a systemic impact, in terms of changing the education landscape as a whole, these Teach For America teachers may very well impact the people who need it the most, namely the kids they're teaching. Having observed first-hand the kind of work that Christina was able to get out of the kids in her class [considered the remedial track] by dint of pushing them in creative ways, I can see that there's a lot to be said for teachers with the glint of idealism in their eye, the energy of youth and [probably quite importantly as well] the knowledge that they just have to keep up the Energizer-Bunny-while-getting-paid-next-to-nothing routine for 2 years, not 30. So, if in the end, a few more kids get turned on to that whole larnin' stuff thing, I guess it's worth it.

I still wonder about what thought process is behind the "doing something meaningful before you start your career" comment, though.