Greg’s Journal Archives
Page 48

< Previous ARCHIVES: Index Next >

ENTRIES ARE ARRANGED CHRONOLOGICALLY. BEGIN READING AT THE TOP.

A Live Journal mirror of this site can be found here, so now you can leave me your comments — even if you aren't a Live Journal member!

Wednesday, November 12, 2008
12:07 a.m.

From Cleveland, The Plain Dealer reports the sad news the Herb Score, the Indians' TV and radio announcer for 34 years, died earlier today. Only 75 years old, the article says, which is surprising because he retired 11 years ago. For some reason, this is making me a little unreasonably sad — I don't suppose I've thought about Herb Score for years, but I keep remembering 12- or 14-year-old me sitting up in my room late at night listening to him announce Indians games on WTAM radio with Tom Hamilton, back during the Indians' good streak in the mid-'90s. The first PD article I linked has a bunch of comments from people describing the same thing — sitting in bed with radios for 10:05 West Coast games and so on — which I guess just reminds me how long it's been since I was 14. In 1997, Sandy Alomar's career year and the year of Jose Mesa in the ninth inning of Game 7 of the World Series, the year Herb Score announced his retirement, I was making a series of recordings of Tom Hamilton's post-game shows by the extremely high-tech process of sticking my little sister's Fisher-Price tape recorder in front of my radio. The shows always contained re-broadcast snippets of the important Indians hits and runs from that game, so there are a bunch of little Herb Score bits that I've been going back and listening to tonight. I won't go all baseball-nostalgic, but it is kind of nice to hear all the old names again and the little pieces of history in progress ("the 171st consecutive sellout," on the way to an eventual 455 that wouldn't end until early 2001). Not to mention my clunkily-made introductions for each game, pre-puberty and usually in a not-quite-full voice because all my siblings were asleep. I also managed to grab a pre-game celebration on "Herb Score night," Sept. 7, 1997. It's a bit long for me to digitize tonight, but Tom Hamilton, hosting the event in front of 42,000 people at Jacobs Field, said a nice thing that fits again tonight: "Folks, nobody will replace Herb Score. We may sit in his seat, we may borrow his microphone, but nobody will replace Herb Score." Apparently (I'm still going through the tape) I managed to record the entirety of the bottom of the ninth inning of Game 7, (even though the me on the tape called it the 10th and the game ended up going into 11), what was supposed to be Herb Score's last inning. I think I go to pretend I'm 14 again.

Friday, November 14, 2008
1:02 p.m.

So... lots of stuff that should have gone before the Herb Score post, except that one was kind of immediate: Going way back, there was an election as I'm sure you all gathered. I would have joined the masses of people making "Yay, something good happened in politics for once!" posts late Tuesday and Wednesday if those weren't the days my cold was at its worst: while Obama was giving his victory speech in Chicago, I was curled up on Car's green cushion thing under my coat and fighting a nasty headache to stay awake. But it was quite exciting to see Ohio go blue for once and watch all the other numbers come in, even though the TV stations seem to spoil things a lot these days by taking 50 billion commercial breaks every hour and pulling in what must be all of Washington, D.C. to make comments on various other things. Also they don't know when to stop with the overloaded CG effects. The results in Arizona, Florida, and California were sad, when I saw them the next day, but everyone else has pretty much covered that in other posts. Friday was the usual Capture the Flag with Stuff, which I didn't play in at all. I think even last semester I only managed one game. This is a bit odd, I admit, because my first two years I couldn't get enough of sneaking round corners and chasing people through Doherty B. But now I just sit sedately in the judges' room and talk with the various other non-players who tend to congregate there, which makes me feel bad about not being physically active when I really do nothing anymore in that line. Based on my camera's poor results in game conditions, I didn't even bother with it beyond the opening meeting. After the game I joined Alisa's party of what eventually became 16 for a trip to Ritter's. Now that Pennsylvania passed a smoking ban, it's so much nicer to sit there! We had to break into four tables of four, but we still overloaded handling capacity to the point that it took hours and hours to get our food. Eventually home and to bed sometime after 3, I think. Just in time, too, because I had to be in Wean 7500 at 11 a.m. Saturday morning for the Microsoft College Puzzle Challenge. We brought back last year's identical Dancing Grues team, consisting of me, Car, Dannel, and Evan, and got assigned for our team room to the newly-discovered excellence that is Wean 4602. (It's the room advertised on the wall with the giant "Copy Room" sign right at the beginning of the 4600 corridor, now converted to a small conference room with a table and whiteboard, and it has an amazing Dutch door that we left open like a service window or something all day.) On the whole, the general feedback was that the KGB hunt we ran the week before was much better than the straight-up Microsoft one, but at least there wasn't a demand for ninja-level origami and art interpretation skills this year. One of my favorite moments early on was a word puzzle with before-and-after clues that required two-word answers that shared a syllable. In response to "Sea officer cooked until just firm" — a lovely image in itself, right? — I give you a first-rate name for next year's team: "Admiral al Dente." There was also a game-show puzzle that Tim would have knocked out in about 30 seconds, if he'd been there, that instead sent us scurrying to the Internet. I solved a nicely manageable thing after dinner that hinged on identifying seven languages from bits of text, realizing that each bit of text was the name of a Harry Potter book title (minus the usual "Harry Potter and") in that language, and indexing by book numbers into the language names to spell out the clue to the answer. We ran into Bengali again, last seen in a disastrous episode at the 2007 MIT Mystery Hunt where the language in question turned out to be Furbish. In addition, the "language guesser" that I knew about online from the Xerox Research Centre Europe managed to identify something correctly as Malay from a sample of three words! The metapuzzle deserves some additional comment. By midnight, when the hunt was supposed to end, we were pretty stuck, having solved a little more than half (maybe up to two thirds) of the 31 puzzles and settled down around 12th place out of the 33 CMU teams. Not much had been happening in terms of progress, and with each puzzle giving one letter of the metapuzzle clue, we were looking at "_W_LV_M___ARR_TBUN_____CETA_GR_." Well, to our diseased minds, this obviously looked like "TWELVE MEN CARROT BUNNY FAUCET ANGRY," so we tried applying the metapuzzle algorithm to those words and got the final final answer of "WOOHAO," which looked like slop. But then we figured that the "TWELVE" and "ANGRY" parts were pretty assured, so Evan grepped the dictionary for all six-letter words starting with W and ending with O. The following occurred: Evan: [Reading:] Whatso... whenso... whomso... woohoo— Me: "Woohoo" is a cool word. [Enters it as the answer.] Computer: CORRECT. Yes. And that is how we ended up in sixth place.

Sunday, November 16, 2008
1:30 a.m.

It's been said that no man is an island. Well, if we take that as correct, I'd say I've at least been feeling very very peninsular over the past 30 hours or so. Today, for example, I woke up a little after 11, saw that it was disgusting outside, exchanged a few sentences with Car, and then read "Harry Potter et le Prince de Sang-Mêlé" until long after the sun went down and I lost track of reality and time. Given the emotionally surexcité ending of the book, this was perhaps a bad idea, but I failed completely at being active or finding any people to do more active things with. Ditto yesterday when the cluster and KGB event couldn't find me enough people to go park-wandering with; instead I came home and watched "Astérix chez les Bretons" on my computer. Either there is some fantastically huge weekend-long party going on that no one told me about and that everyone's gone to, or I've been singularly bad or lazy in my choice of timing. Which means, in the second case, that as soon as I get crazy-busy again with schoolwork, then everyone will be wanting to track me down to do stuff. In general, I've been increasingly under the impression that Car, Tyler, Pyxy, and I are only connected by the fact that most of us happen to sleep at the same street address most of the time. Five days out of six this week I've come home to an empty house, dark and cold in this season's typical weather; we all must be measuring the existence of the others mainly by the number of dishes piled on the kitchen counter and the state of the shower as we wake up in the morning. Not that this is anyone's fault in particular, given that it's that panicked time of the semester for all of us, but I suppose I'm just noticing more this week that our house never turned into the social hangout we all seemed to hope it'd be when we moved in. I guess it's been a long time since junior and senior years at Case, when I could just leave my door open and have about a 90 percent chance of attracting people or diversions or dinner plans or impromptu movies. Perhaps I'm a pathetic old grad student, but right now I kind of miss that.

Tuesday, November 18, 2008
2:46 p.m.

Linguistics post today, since I spent several hours Sunday talking about language-related stuff, and that made me happy: The best sentences ever always come from The Tartan's copy before (and sometimes after) we get a hold of it in the copy cave. Sunday afternoon, we were first confronted with the description of a certain novel, "one of those epic books so famous everyone thinks they've read it but haven't." We had to put that one up on the board: the usual singular "they" silliness in the first half, followed by an authorial change of heart, I guess, where the writer decided that the pronoun really should be plural after all. It took us several rounds of proposed fixes before Scott finally hit on the relative best one: "one of those epic books so famous most people think they've read it, but they haven't." We also got a "less"/"fewer" question, which meant I pulled out the fun example sentence I'd been saving for a few days: "There are ____ obvious reasons for wanting to do that." I think the other editors' first instincts were "fewer," but it took them (as I expected) almost no time at all to realize that either alternative can work, just with different meanings — and, for a bonus, you can even have both if you put them in order "fewer less." We had like seven people in the copy cave at that point, and not enough work to go around, so I amused myself by drawing a simple tree-adjoining grammar for the sentence. Elementary trees are ridiculously hard to visualize nicely in text, but the main idea is that you have two "less" trees: one that adjoins to a countable-noun NP and makes a larger countable-noun NP, and one that adjoins to an ADJP to make a larger ADJP that eventually gets inserted into an NP of any type. And finally, from the land of dialectology, comes a piece of information apparently about New Jersey accents in which some people will replace at least word-initial "tr" with "chr." (It makes more sense in phonetic terms as [t] going to [tʃ].) If so, it must be pretty subtle, since I had a whole conversation about trains with someone who apparently does this before he mentioned it, and I hadn't picked up anything. But I guess most phonetic features like that are hard to find unless you're specifically listening for them: otherwise, you hear the intended word and your brain ignores that there might be slight variations in the actual articulation. And now back to your regularly-scheduled Tuesday.

Saturday, November 22, 2008
1:38 a.m.

We all know, I expect, the famous quote: "There is a tide in the affairs of sinks, which, taken at the flood, leads to a nasty half-hour of cleaning everything up." This is of course nothing more than my embellished way of saying that the faucet on the kitchen sink started leaking today from a little hole that happens to be at the back of it, covering our counter in water from one end to the other, and that I was the first person to discover it when I went downstairs to throw together bread and peanut butter for what I thought would be a quick lunch. I don't know exactly what happened to set this off, but it's probably approximately fixed — at least for now. And our counters have had a good rinsing. I say, if you ever wander back into a seventh-grade science class and hear the teacher going on about surface tension, you should put in a good word for it on my behalf: it's the only reason we didn't have a giant mess on the floor as well. Other developments in things too, but I'll save those for a time when it's not so late.

Tuesday, November 25, 2008
4:26 p.m.

I think — although now of course it's been long enough that I can't be entirely sure — that the major "other development" I promised last time was related to the fact that someone — that is, some company — is finally interested in me enough to want me to come out for an interview! That sentence deserves an exclamation mark because this is something that's theoretically been in progress since mid-September, when I talked to these people at the TOC and had a non-technical first-round interview on campus the next day. Eventually, after some amount of worrying on my part, I got asked for a code sample, which I sent in, and the satisfactory examination of that eventually led to an invitation to this second round. So if you don't see me next Thursday or Friday, it's probably because I'm in Boston. I'm excited, since I've never had such a thing before ever, and I really want it to go well. The most interesting part of my six-hour interview at the company is that they want me to give a 45-minute technical talk. Overall, I think that's a good thing because Alon's been saying that I've come to give pretty good talks recently, and the one I'm adapting for the interview is based on the one that won me $100 for honorable mention at the SRS in September. Also good because I feel much more at home in a discussion about MT than, say, being asked to write recursive C functions on paper on the spot for mucking with some data structure I haven't dealt with since my sophomore year of undergrad. The company itself would be an excellent fit too, given the things they've worked on and published in the past few years. Otherwise... I probably wanted to say something about Lemmings, a cute little video game from 1991 that I saw for the first time at the KGB/C-Club "classics" night on Friday. This is going back far enough in video game technology that I can handle it and find it fun, and that's before you get to a surprisingly good remix of "London Bridge" as the music in one of the middle levels. I managed to track down a fairly poor web version after a quick search Saturday, but the popularity of the game makes me expect I can find something better if I try again. Something's a bit wonky with the academic calendar this year, so we had our last issue of The Tartan Sunday instead of right before the last week of class. The evening copy staff was out in full force and broadsheet ran ridiculously ahead, which meant that we had time for a game of Scrabble (which I won at the last moment by getting 11 with an "I" as my last tile) and an interesting dart-gun battle à la 2005. Well, we actually split into two groups, took up positions on opposite sides of the room, and then fired at the wall behind the other group instead of at the people themselves, but it kept an even flow of darts back and forth and quickly evolved into a target-hitting competition where we each tried to get as many darts as possible as high as possible on the wall to keep them out of the reach of the opposing group. This continued until Andrew called us out for the usual end-of-semester business in the main office, from which I didn't return until after 2 a.m. Yesterday it rained most of the afternoon and evening, and I went to the AUO concert. I had been kind of commissioned to be the official photographer after the AUO sent an e-mail to the rest of the orchestra looking for a photographer, and Keith (who's a member) thought of me and forwarded the message. I took about 140 shots, I think, and felt overall pretty good about them. Of course, that may change once I pull them off the camera and have a look at the full resolution, but I'll state for the record in advance that the low light was bothering me less than it usually does in these kind of situations. Today I can do no better than to share with you this information about the island nation of San Serriffe, which I expect at least some of you will find amusing. My next major goal for the day is to get word alignment running on a parallel French–English corpus of 9.8 million sentences. This should have been started Saturday, but I had a pile of trouble getting the English to parse correctly and eventually decided just to give up and take what I've got so far, so hopefully that'll be underway soon.

Saturday, November 29, 2008
12:07 a.m.

Aside from various political affairs and a shocking disrespect for the environment, there are few behaviors of our U.S. citizenry that make me less proud to be an American sometimes than rampant consumerism. Today we have decided to cross another such threshold. No longer are we content to spend nine hours lined up in parking lots in the top-priority pursuit of things when we ought to be at home playing Scrabble or making weasel faces at our siblings over the last bottle of sparkling grape juice, but we now find it necessary to burst into stores at the stroke of 5 a.m. with such ferocity and lack of concern for anything else that the employee unlocking the doors gets trampled to death. Revolting. And when the police told the shoppers they had to leave because the store was closing, they got upset. I say, is Boxing Day in Canada or Europe like this?

Wednesday, December 3, 2008
9:49 a.m.

I'm sick! Again! With the same thing I got a month ago that took me more than two weeks to get over! This is of course marvelous timing, as many of you know from your own experience, because this is the last week of class for the semester. So not only did I have a conference paper due Monday, a machine learning paper due today, a poster to present for the same class this afternoon, an interview to prepare for and a talk to give Friday, and a final exam next Monday morning, I also have a translation system to build and get ready for the WMT eval that runs all of next week. If I manage to bring half of these things to their satisfactory conclusions, it will be an amazing stroke of luck; right now, I just want to go back to bed and sleep for a week. And probably rip out my lungs so I don't stay awake with a spasm of coughing every few minutes. In conjunction with part of the above, I should announce that I'll be appearing in the Newell-Simon atrium today from 2:30 until 5:30 (or something like that... basically all afternoon) standing next to a poster headed "Automatic Selection of Grammar Rules for Syntax-Based Machine Translation," which if it were due right this minute would consist entirely of the title and four unlabelled diagrams. The idea is that each of us in the Ph.D. machine learning class has to talk to all interested parties for three hours straight about our class projects, and there's some kind of voting for the best one. In class, it was set up as a sort of popularity contest along the lines of "Get all your friends to come and you might win!" — so you guys reading this should all come and subvert the system by voting for the project you actually think is best. Bonus: if you get there at the right time, you might get to see me hack up my innards all over the professor, unless my supply of cough drops somehow proves more potent today than it did last night.

Sunday, December 7, 2008
11:45 p.m.

It seems reasonably assured at this point that I'm going to the MIT Mystery Hunt in January again. It also seems reasonably assured that the state of airline prices will force me to rent a car and make it a 12-hour road trip instead. This is actually not a bad thing, since I like intercity road trips, can rent cars (since my last birthday) without paying the "you're young, so we don't trust you" fee, have a place to stash the car for free in Boston during the hunt, and am set up for massive savings over plane fare if I can find more than one person to come with me. I think Drew's already confirmed for car-sharing, but it's standard for a car to fit for people comfortably. So if anyone else wants to make it a cheap trip to the Mystery Hunt, let me know soon! The plan would be to leave Pittsburgh around 9 a.m. on the morning of Thursday, January 15 and basically spend the day driving. The hunt starts Friday at noon, and then we'd hope it ends by Monday morning so we can spend that day driving in the reverse direction. This is at the end of CMU's first day of classes, and Monday's Martin Luther King Day, so we'd be just missing the Thursday and Friday. Again, let me know pretty soon if you're a confirmed passenger so I can officially rent the thing and get you an approximate price.

Thursday, December 11, 2008
12:34 p.m.

The last time someone asked me the Live Journal five-questions thing, it was April 26, 2006 — well, early in the morning of April 27, technically — which was the night I slept in the Tartan office after working on my Grammar Formalisms project (and the answers to Dan's five questions) until 4 a.m. But it was still fun, and now the questions meme has come up again, so I will present it to you again! This time the questions are by Kempy, and my answers are coming out so incredibly long that I'll have to post them separately over the next few days as I write them. (1) What was it like growing up as you? Describe your childhood and in particular, your relationships with your siblings. At an abstract level, I probably had the same sort of childhood a lot of CMU- and Case-style nerds did. Early and avid reader, obsessed by numbers, used to scratch long division problems into the dirt at recess, smattering of QBASIC from the age of about eight — those sorts of things. From the beginning, I was a small, shy, quiet kid who liked to be precise about details and facts. Outside of the typical nerdy stuff, I think a lot of the way I grew up was influenced by where my family lived from when I was three until I was 11, which was a small town in Michigan that the 1990 census pegged at about 6000 people. (We used to have walking field trips from school sometimes because places were close enough, and I guess people were less paranoid.) Behind us was a big open field, until they started building in another neighborhood, and at the end of our little cul-de-sac was a park and then some woods with a creek in it. Between the two, my sister and friends and I would disappear for hours and hours, running off to wherever our current settlement of choice was. The leftover construction material from the new development meant we could "build" our own little houses and shelters and things; led by a friend a year older than me, a group of about six or eight of us built a wooden boat and sailed it in the creek. (There's video of this, which is amazing to see after all these years; I was about seven or eight at the time.) So we had this gigantic playground of creativity and fun, which I know my younger siblings didn't get because we'd moved to a more suburban place by the time they were old enough to care. Then there was all the adventures my sister and I used to cook up for ourselves. I'm the oldest of four rather paired-up kids: my first sister is two years younger than me, then there's a gap of four years before my next sister and brother are only a year and a half apart. (Translation into current ages: 25, 23, 19, 17.) So growing up this was always the older two versus the younger two, but being the oldest meant that I could do all sorts of fun little-kid things way after I should have been able to. My sister and I were almost inseparable and had most of the same friends in common all the way through high school. Chris and I used to make up all kinds of long-running fantasy worlds when we were at home. As just one example: at some point we created our own pretend TV network, filled a broadcast schedule full of shows, performed them, and recorded a cassette tape full of commercials to play for actual commercial breaks. The more I look at these kind of things in retrospect, the more I feel proud that I did all this very silly and very creative stuff up to an embarrassingly high age, because I think it's been much better for me than a childhood of Internet and video games and cable TV would have been. (My family's never had cable, and we didn't have the Internet at home until I was 14.) I find myself right on the cusp between an older group of people befuddled by personal electronic gadgets and a younger group of people unable to pay attention to any one thing for more than 15 seconds. I went off to college when my brother was still only 10, so I feel like I missed about half of the younger two siblings' growing up. I feel like I know them a lot less as people than I do my other sister. People with younger brothers and sisters can perhaps confirm that, when you're 12 and they're six, they're a blot on the landscape and an annoyance to anything you want to do without them, but at some point you cross some age or wisdom threshold and reconsider that, when you were six, you wanted pretty much the same thing they did and managed to get off without an older brother bossing you around. This still continues: every time I go home, I keep finding myself vaguely surprised that Katie and Andrew are full-functioning mostly rational adults instead of the annoying eight-year-olds I seem to remember. This is especially true in the case of my brother, who has a girlfriend and a life and a set of personal hopes and future expectations I know very little about. He should also certainly not be allowed to be an inch taller and 15 pounds heavier than me...

Friday, December 12, 2008
9:09 a.m.

I've been up since 5:30 for some WMT evaluation work, but I have the answer to Kempy's next question ready to go, so I thought I'd post it. If you missed the background for all of this, you may want to have a look at my previous post. (2) What is it like being a grad student at a different school than the university you attended for your undergraduate degree? How would you compare your own experience with the experiences of those around you who attended Carnegie Mellon as undergraduates? Ooh, another good one. There are various partial answers to this scattered throughout my journal archives, most notably in response to a question Alisa once asked about what I missed most from Case. The biggest single difference, I think, is what the first year is like. As an undergraduate freshman, you come into campus housing all revved up for this great college experience you've heard about most of your life and seen on TV and all those things, and everyone's pre-programmed to meet everyone else. I don't think it took me more than the first few weeks of class, for example, to meet the people who'd be my closest friends for our entire four years. None of that really happens in grad school. I don't know if it's because everyone's already done the "college thing" and is looking for a mature or formalized social life, or if they've all gotten married to their undergrad significant others and just want to spend most evenings quietly at home, or what, but I know very few people who seemed to be socially happy during their first year in grad school. I know for myself that it took a semester or more before I felt like I wasn't just some appendix flapping uselessly next to some cohesive and complete social unit that didn't need or want me, and the result of that (and living alone in a one-bedroom apartment 30 minutes' walk from campus) had major effects on how I think and what's happened to me since. I had a lot of time to do nothing but analyze my own thoughts, which really turned out to matter a whole lot just under two years ago. There are some other little things, like my database of terms and names for things from Case being slowly overwritten by CMU stuff. When I first got here, I had the worst time introducing myself as "from The Tartan" rather than "from The Observer," and in copy editing I always had to look up or ask someone else the official names of Student Affairs or Housing and Dining Services. Now it sometimes goes the reverse — I'll talk about a "McConomy movie" when I really meant to say "Strosacker," and I couldn't tell you anymore what we had at Case instead of "Facilities" or "FMS" here. My experience as a CMU grad student and your experience as CMU undergrads? Mostly different because of academics, as you might expect. I can manage to live a whole lot like an undergrad to a first approximation, which I definitely do, but no matter how many late nights I spend in Donner or the cluster, we still live in two very separate academic spheres. Such a huge part of my undergrad experience was with people who had that same common foundation, and at the level that we were all "the engineers," we were at Case for the same reason and were going through the same process at the same time. Since the LTI has basically no undergrads and I don't really have close friends in my own department, that's no longer the case, and I'll admit (and this journal will show) that it's given me some very difficult periods of feeling very stupid when everyone else around me spends six hours talking about kernel thread creation. Beyond academics, another major (and frustrating) difference is the work schedule. Grad students are simultaneously incredibly flexible — little class, mostly self-defined hours — and ridiculously locked in — "Just because classes are in session doesn't mean research work isn't getting done" and "Spring break is really not for you" are both things that have been told to me since I've been here.

Saturday, December 13, 2008
1:47 p.m.

Apparently these long rambly answers are coming out OK, because I got an explicit request last night for more. I was actually thinking, on my way to campus yesterday, that a pretty good first attempt at an autobiography (or at least some related project) would be to just take my answers to each one of Kempy's five questions and expand them into chapters. They're already almost long enough, and some of them I've even cut down to keep them readable. Anyway, here's the next installment: (3) What inspired your love of languages (and grammar)? Suppose you could speak English, Thai, and French perfectly fluently. What other language would you want to speak fluently? Would you rather speak that language fluently or speak that language and another, with less proficiency? Honestly, I'm not sure if I remember a time when I didn't want to know French. The origin of this is a little unclear, but I expect it's some combination of being about to get Canadian TV in Michigan and the fact that my mom took two years of French in middle school (but had forgotten all of it by the time I was around and asking for words). Then I had a string of teachers in elementary school who were very careful about teaching us the mechanics of punctuation and such, which somehow I took a liking to. (Stuff I wrote at nine shows that I knew what a semicolon was, but only partially how to use it.) So I started taking French at the first available opportunity (ninth grade), but it was probably my 11th and 12th grade English teacher who really got me into writing. She had a master's in composition and always looked at our essays from a stylistic point of view. She really liked my writing — which I think was the first time I had any clear indication that I was any good at it — and, I'm sure, made it a lot better. In 12th grade was when my love affair with journalism started, and I became a copy editor pretty much right from the beginning. The kind of detail-oriented word nerds who flock to copy editing are wonderfully fun people to be around — our axis of similarity is so different from the KGB/cluster group that it's often a striking change to be related also to the liberal arts world. The language technologies thing, which of course has gotten much stronger since I've been at the Language Technologies Institute, grew mostly out of me wanting to go to journalism grad school but not wanting to give up that CS style of work I'd done as an undergrad. (And, you know, finding that journalism grad school cost $65,000, while they'd pay me a bit to come here). I had found I liked CS much more when it was applied to something I cared about, and here I've pretty well mapped out that translation is something I care about and find interesting. But before grad school I had never taken a linguistics — much less a computational linguistics — class. I had always wanted to take phonetics at Case, but it conflicted with The Observer's production night and I was too entrenched in the newspaper to give it up. Since I've been at CMU, I've gone a lot further in picking up little interesting bits of information about various languages, even though I speak very few of them, which I think gives off the impression that I'm more of a compendium of human speech than I actually am. A waiter at the Rose Tea Cafe, for instance, once saw me puzzling over a Chinese newspaper and asked me excitedly if I could read Chinese characters — the answer is no, just digits, dates, and "de," which means I can kind of find where some of the compound noun phrases are. The comic strip "Frazz" posed the question, maybe five years ago, "Would you give up a quarter of your English vocabulary in order to have equal fluency in some foreign language?" That's a major stumper for me. I like having a decently large supply of English words at my command, but it also frustrates me manifestly in French when I don't have the same readiness in sentence construction or conversation that I have with a native language. Listening to a natively bilingual person effortlessly switch between perfect control of two languages makes me really jealous. So far I've focused on knowing just one foreign language well, which has gotten me mistaken for a Belgian in Canada and which I hope I can always keep up. Beyond that, I'd really really like to get to at least a conversational level in Thai. Learning by myself, though, makes it quite hard and also makes me quite nervous when I suddenly have to jump from my own very untested knowledge to interacting with native speakers. It's situations like these, when you have just a glimmer of knowledge, that things are the most awkward, I think, so I feel like I'd rather focus on decently good knowledge of a few chosen languages. I think that would make me feel more accomplished than knowing things like "Where is the bathroom?" in 20 languages — if I ever have money and time to travel, I could try to pick up those things on an as-needed basis in the months before a trip abroad. In the ideal world of me being already fluent in English, French, and Thai, there are a few other languages that I'd be the most tempted to try. German's high on the list because of an exchange student friend from high school, and Chinese will be very important in the next 50 years. I currently live in the wrong part of the country to have many opportunities for interaction in Spanish, but as far as foreign languages in the U.S. go it's probably among the most useful.

Sunday, December 14, 2008
5:00 p.m.

My answer to today's Kempy question may get away from what was actually asked a bit, but no one has yet accused me of putting them to sleep from having read my entries during an important final exam study session, so I shall continue being verbose. After all, you only have one more to put up with after this one! (4) Is there a talent or a skill that you don't possess, that you wish you did? These questions are leading from each other pretty nicely, I see. But I'll assume all the language desires from last time as a given and work here from a slightly different angle. Alert readers of this journal may know that I had a rather difficult year last fall and spring from mainly one personal angle: I felt very intensely that everyone else had all these wonderful skills and opportunities that I didn't have, that I was hopelessly a failure at learning any of them, and that whatever irrelevant things I had to offer from my second-rate background would never get me through grad school, into a job, etc. and so on. This is obviously not true, as a few excellent people kept telling me then and as I've been trying to rediscover again by some changed habits. One of the things I was most worried about during that time was that I was too slow to pick up any new thing — that everyone else could sit down at a video game or a new programming language and have it mapped out in their brain and be succeeding at it in about as much time as it took me to say "OK, so I've got this control stick, right...? And it does what again?" Again, further reflection tells me that this is probably mostly not true — and if it is, it's because I've been forgetting to normalize for things like background and personal interest. I mean, if I didn't go to community college and computer camp and fancy schools when I was younger, I won't know the stuff that all these other people learned there, and if I've never really ever been a video gamer, it's not likely that I'll pick up one of these dizzying modern games and do very well at it. But one thing that all these people I compared myself against had in common — which will get somewhat back to this actual question — is a strong intellectual curiosity and a strong will to force themselves to just do things. Example: I have over 8000 photos on my hard drive; the difference between that and this is that I haven't yet sat down and forced myself to learn good PHP and Photoshop to make something like that happen, which is what I'd have to do if I wanted to end up with a nice photos site like that one. The general pattern carries over into a lot of things, even in the non-personal realm: I could learn and accomplish more in grad school if I spent more time honestly trying to make the right kind of progress. The trouble is in forcing myself to do things that I kind of rate as "not fun" or "annoying" or "way beyond me," and the people I wish I could be as talented as don't seem to suffer from that base motivation problem. If I could teach myself to stop caring that, yes, I may be plodding through something at the age of 25 that my friends did more easily at half my age, and could focus my efforts more into the things that I both want and need to get done, I feel like I'd be a lot better off. It's that particular skill that I feel like I had before but have lost in recent years. Given that, a lot of other skills I want should come more easily. I do wish I could pick things up faster, but I feel like even slow learning would be OK if I could approach it with a better mindset. There are some specific things I would love to be able to do. If you rewind the tape of life to last fall, you'll find me desperately wanting to get better at ITG and learn Photoshop and fancy web coding stuff beyond my basic HTML. I've pretty much always sucked at sports, which doesn't bother me now like it once did, but I still find myself only a mediocre runner: better, I suppose, than people who have never really tried it, but worse than anyone who has. Since I was at a choir concert last night, I'm reminded that singing and music in general are two easy things to point to that I could have developed skills in but didn't. I took two years of piano lessons when I was nine and 10 (and mostly didn't like them at the time because my mom forced them on me and my sister), which means that 15 years later I can only (slowly!) read sheet music and play some scales. Then I was in choir for three years in middle school but dropped it going into ninth grade in favor of computers and French, which means I remember some simple pronunciation rules but am terribly shy about singing in public. Both of these are right at that awkward stage I mentioned with languages yesterday: knowing just enough to know you can't do what you want with the thing. For more than two years now I've had this sort of pipe dream of writing a musical based on the cluster (I think I got the idea from Matt Aument in KGB), but I've never gotten farther than partially complete parody-style song lyrics — actually learning to play them in order to even make preliminary voice and piano recordings would take me months and months.

Friday, December 19, 2008
11:24 p.m.

I am always doing this: getting partway into some nice routine, then disrupting it once for what should be some small reason and consequently breaking it entirely. So now I see that I've left the fifth Kempy question hanging for five days. It may have something to do with the fact that I think it's the hardest to answer out of the five, so I've been procrastinating editing my final answer for it. It's not that I can't think of what I want to say... more that, no matter which way I put it into words, it comes out sounding a little off. My apologies for any remaining irregularities in what follows. (5) Everyone is good at something. At what thing or things do you believe you excel? This is the part where we tie together the answers of all the previous questions, I think. I've been told my French is quite good and that I have a decent knack for "language stuff" in general like learning new ones, editing, writing, and so on — which I think is partially true. Which, you know, is probably a good thing since I'm hoping to make my career in machine translation or at least computational linguistics. At some point, I started noticing that I often think to myself in complete sentences, which seems to translate into being able to coherenly verbalize concepts that other people are explaining, and that can also be a useful skill. For a long time, I thought that one of my strongest points was that I've had a decently wide variety of interests and experiences over time rather than setting myself up as a one-sided expert in a particular small area. There are piles of people, especially here, who are better coders than me. There are a number of better copy editors and journalists. There are people who speak better second-language French, and probably anyone who's ever tried has gone further in Thai than I have so far. But if there's ever a situation — a news room or a copy desk would be the best example I've come across so far — where disparate things have to be pulled together at once, I kind of feel like I can hold my own there. I've always had a thing for collecting and saving random facts — I seem to retain a good amount of knowledge, from things I learn or do, that's easily expressible as factoids — and there have been some fun times when I pull something apparently unexpected from my store into an unrelated context. The people at the Mystery Hunt in 2007 seemed to give me instant credit as an electrical engineer when I threw out some things I remembered about 7400-series logic chips. At that same hunt, I think I was the first to spot that a given list of words contained those that could be Caesar shifted into other English words because I'd just been looking the day before at an old MATH 408 assignment where we had to find them. I really like being able to do stuff like that: put things together from different backgrounds and feel like I'm making some useful contribution to the current problem. If I started that last paragraph with an implicit assumption that I no longer find that to be one of my strongest points, it's only because I may question its general usefulness and also its uniqueness. As far as usefulness goes, you could make a strong argument saying that what someone working in software really needs to be is just as good a coder as possible, and who cares if he can give you the word for "tea" in seven languages or not. In that case, I'd be less useful because some of the energy that I could have spent on learning Java I spent instead on taking an ASL class my senior year at Case. This is one reason why I really enjoy newspaper work: that's definitely one job where you need to pull in a whole bunch of different knowledge sources, and the fact that I can edit (say) SciTech articles with some idea of what they're talking about sometimes is a good advantage. It's also fun, never knowing what might come up next. Along the uniqueness dimension, it's all well and good to say "Whoo kids, look at me! I'm good at a lot of things!" — but, honestly, isn't that the case for quite a lot of people? I knew people at Case who double-majored in electrical engineering and classics, or computer science and communications, or chemistry and theatre. Or people here at CMU who are equally superior and adept at discussing C compiler optimization and obscure art or novels. There are, in the end, very few truly one-sided people, although some obviously are more concentrated than others. Is there some difference in the way I accumulate or make use of my varying skills compared to other multi-talented people? That's for other people to decide, perhaps. I do feel like — if you can stretch the definition a bit and call this a skill — that I have a certain nice moderateness that tends to work well with other people of varying viewpoints or backgrounds. One of my long-standing complaints, if you will, against CS-type people is that we're all so overwhelmingly opinionated. If someone thinks that ML is a well-constructed language and Java has some clunky quirks that aren't so nice, he won't express in moderate terms a personal preference for the one over the other for his own applications; he will unequivocally tell you as fact that Java is "doing it wrong" and that ML is "right." I find this highly annoying (whether it's in the domain of programming languages or salad dressings) and try to avoid things like that in myself, which could arguably entail the skill of not making people hate me. But again, that's for the world to decide and not for me to posit with no external evidence.