academia | advice | alcohol | American Indians | architecture | art | artificial intelligence | Barnard | best | biography | bitcoin | blogging | broken umbrellas | candide | censorship | children's books | Columbia | comics | consciousness | cooking | crime | criticism | dance | data analysis | design | dishonesty | economics | education | energy | epistemology | error correction | essays | family | fashion | finance | food | foreign policy | futurism | games | gender | Georgia | health | history | inspiration | intellectual property | Israel | journalism | Judaism | labor | language | law | leadership | letters | literature | management | marketing | memoir | movies | music | mystery | mythology | New Mexico | New York | parenting | philosophy | photography | podcast | poetry | politics | prediction | product | productivity | programming | psychology | public transportation | publishing | puzzles | race | reading | recommendation | religion | reputation | RSI | Russia | sci-fi | science | sex | short stories | social justice | social media | sports | startups | statistics | teaching | technology | Texas | theater | translation | travel | trivia | tv | typography | unreliable narrators | video games | violence | war | weather | wordplay | writing

Wednesday, July 20, 2016

Every error is an error of substance

I've unlearned all of the accuracy-boosting skills I once learned in sixth grade about typing. I've turned into a truly terrible typist, in large part because I've picked up the nasty habit of trying to correct my errors as I'm typing. If, say, I'm typing a passage from a book, I can read faster than I can type and I get ahead of myself. I hit the delete key in the middle of words but miscalculate the number of letters to go back, leaving half-formed words. A refresher typing course would not be a bad idea. It seems especially painful to be writing a dissertation about error correction and then to start erring left and right, enacting the very thing I'm writing about.

Except that I
Slate article on spell-checkers

Incorporating user data is a huge step in the right direction for Word, but the process is still sluggish compared with search engines. Google and Live Search generate dictionaries that approach real-time models of language. In a fascinating paper (PDF), two Microsoft researchers explain that a stream of previous search queries can be used to maintain an up-to-date lexicon capable of correcting a high percentage of mistakes, even when 10 or 15 percent of your searches have errors. This purely statistical approach is much timelier than any involving human editors and has far fewer biases. When it comes to fixing errors, the researchers write, "the actual language in which the web queries are expressed becomes less important than the query-log data."


cites Louis Menand's 2003 article about Microsoft Word
Try to prevent Word from doing that blue thing to whatever it recognizes as a hyperlink. There is undoubtedly a way to reset this, but it is deep within the bowels of the machine, guarded by dozens of angry pop-ups. Microsoft wants you to go on the Internet.

but what if this feature had a useful function, as the hyperlinks in Chris Wilson's article (including the one to Menand, in a nice outside-the-Washington-Post-company-hyperlink that we should see more of). There's something vaguely Andy Rooney-ish about this list:
Attention to the new demands of electronic media informs almost every chapter of the new “Manual.” There are discussions about (besides citation) preparing electronic publications, editing and proofreading onscreen, and electronic-publishing rights and permissions. The authors are sensible about these matters; they’re aware that this is an area very much “under construction.” In all departments, in fact, the authors allow themselves plenty of wiggle room, quoting a passage from the 1906 edition: “Rules and regulations such as these, in the nature of the case, cannot be endowed with the fixity of rock-ribbed law. They are meant for the average case, and must be applied with a certain degree of elasticity.” This is modest and becoming, but it is beside the point. The problem isn’t that there are cases that fall outside the rules. The problem is that there is a rule for every case, and no style manual can hope to list them all. But we want the rules anyway. What we don’t want to be told is “Be flexible,” or “You have choices.” “Choice” is another of modern life’s false friends. Too many choices is precisely what makes Word such a nightmare to use, and what makes a hell of, for example, shopping for orange juice: Original, Grovestand, Home Style, Low Acid, Orange Banana, Extra Calcium, PulpFree, Lotsa Pulp, and so on.

That is, why make the list if you're not going to comment on the nature of lists and corrections: they're both methods of containing options that end up proliferating them. It's the entire modus operandi of Entertainment Weekly.

Why is error correction linked to nostalgia?

Labels: , , ,

Monday, July 18, 2016

Amazon, quality control and brand erosion

Amazon is the everything store for me and so many people, which is why it's such a problem that you can't trust their product listings to match what they say they match, or to distinguish between different editions or even different translations of a book.

It's a total mess. They aggregate the reviews as if it's the same book, so people criticize one translation in a review, and it's interpreted as a critique of a different translation!

And of course that means it's a gamble to order a used book... often Amazon hasn't even told the seller which edition you thought you bought. Then you're faced with the prospect of punishing the seller for something that's the platform's fault.

More broadly, Amazon has had a problem for many years with the fidelity of suppliers' items. You can't review a partner seller Amazon is in bed with, or report that they fraudulently pulled a bait and switch, only return the item at your own time expense. So plenty of these second-party sellers build up a few reviews by selling the real thing, often through the Amazon trusted reviewer program, then shift to knockoffs. Amazon makes it very easy for them to give Amazon extra money for this process, then makes it hard to resist that you're being sold a lower quality item than the one on the website.

I think it's a mistake to let their brand erode like this for nickels and dimes... I stopped buying much on eBay when it became clear that their PayPal fraud guarantee was designed to not pay you back even when eBay and PayPal admit your money was stolen. I'm much more cautious about buying on Amazon these days if there are multiple editions of a book, or there's any hint of fraud in the comments (which is startlingly often).

FWIW, the bookstore Schoenhof's in Cambridge MA is an excellent place to buy, and ask about, books in languages besides English. (I recommend reading Harry Potter in translation--it's a great way to improve at the intermediate level!)

Labels: , , , ,

Tuesday, July 12, 2016

Imagination and its poisoners

I don't like arguments that seem to assume there is some kind of analysis that only deals with statistical data without additional reasoning, interpretation and imagination.

See the old debate between Ta-Nehisi Coates and Jonathan Chait. I think Coates, for all his great points, is too quick to insist that the debate stop using speculation and stick to research. It's not that he's wrong about the value of research, but that he's wrong that it would do much to clarify his and Chait's debate. The deep problem is not that they have much specific disagreement, but that they are playing for different teams. I wonder if it is Chait's hitting close to the bone, rather than his missing the mark, that raises Coates's ire; cogent arguments by someone who disagrees with you are welcome when you feel you can safely allow the other side to score points, but they are panic-inducing when you feel you cannot give ground. Then, entertaining thought exercises feels unsafe.

In a response to Stephen Pinker's The Better Angels of Our Nature and its argument that violence has declined, Nassim Taleb recently commented that "You can't easily extract statistical information from fat tailed data". I think that phrasing suggests a basic misunderstanding of statistics and data and reasoning... there is no pure information devoid of context that can be extracted from data. (And Pinker stays away from misleadingly clean math in his argument for a decline in violence.) There is no point, in collecting evidence, at which you suddenly have enough data to justify interpretation.

Think of the decisions we make in our lives and how little data we have available. Can I point to rigorous statistical studies that tell me my daughters will have a better outcome if they go to our local public school, where the administration and the teaching staff strike me as incompetent and unprofessional, versus if she goes to her actual current school, which is well run and staffed by people who seem, anecdotally, to love children? Of course not. Is my sense of confidence that this is a better choice a worthless illusion? No, it's not--I could well be wrong, and I should push myself to reduce bias in many ways, but my confidence is built on an understanding that is at least somewhat sound. Taleb would probably acknowledge that if asked in isolation, but he doesn't recognize that he precludes this sort of soundness when he dismisses the value of cogent reasoning from limited data. (Something which he himself does plenty of.)

There is no magic shortcut we can use in our analysis, statistical or not, that avoids the need for reasoned and impressionistic interpretation, nor any method that avoids the possibility of our reasoning leading us to being wrong. There is only the actual quality of our reasoning, the carefulness of our interpretation, the multifaceted understanding that we bring to bear, the clarity of our process, our willingness to consider conflicting evidence, and the intellectual honesty to keep score without regard to whether our side or the other has scored each point.

I have read plenty of work by those who bullshit the reader and lie to themselves to promote a point. They have certain consistent characteristics that act as tells: they use ridiculing phrasing to belittle opponents, even when mentioning something that is perfectly normal; they treat their perspective on complex topics as certain, rather than provisional and tentative; they treat their conclusion as so obviously correct as to shame those who have denied it. These are signs that thinkers are not merely our humble allies in processing the complexity of the world, but instead need to enlist our support in nursing their grudges and giving them back the power they deserve.

Pinker is not free of all of this; I would have preferred more recognition of his limitations and the limitations of the data available to him, and more humility in his asides about culture and identity politics. But Taleb's writing shows all of these signs of dishonesty and resentment. (Gary Taubes is another.) When I read his writing--I, one reader, because there are so many other valid readings--I perceive viscerally that the writer feels cornered and that there is danger in the air. An ally in interpreting the world, secure in the soundness of his reasoning, doesn't write publicly that "Problem with Pinker boy is that he made statistical claims without knowing what he is talking about." A bully, who prefers being respected to promoting depth of understanding, does.

When I read polemical writing--which the world needs!--I try to ask myself if the writer is trying to grow my imagination, or constrict it; to seed ideas plentifully, or to poison them selfishly. It can be very hard to notice this through talented prose; read Michael Lewis with this lens, and his romps through tricky topics look quite different.

Labels: , , , , , ,

Friday, July 08, 2016

The Super Mario principle

The other day I saw a group of teenagers playing 3ds Smash Bros against each other. They said they never, ever use the 3d and just leave it off. (Actually a bunch of them just bought the bigger screen ds with no 3d at all.)

Just a reminder that technical advances in the presentation of entertainment are not necessarily a big part of its appeal, compared to UX design, networking, the social layer, simplicity, and cuteness.

Cf. Minecraft, Scratch, SMS, Candy Crush, Angry Birds.

My daughters LOVE playing Super Mario Bros. on an old NES--a game more than 30 years old.

Labels: , ,