Tag: computing

Gender and the Metropolis (algorithm)

I’ve always heard of the Metropolis algorithm having been invented for H-bomb calculations by Nicholas Metropolis and Edward Teller. But I was just looking at the original paper, and discovered that there are five authors: Metropolis, Rosenbluth, Rosenbluth, Teller, and Teller. Particularly striking having two repeated surnames, and a bit of research uncovers that these were two married couples: Arianna Rosenbluth and Marshall Rosenbluth, and Augusta Teller and Edward Teller. In particular, Arianna Rosenbluth (née Wright) appears to have been a formidable character, according to her Wikipedia page: She completed her physics PhD at Harvard at the age of 22.

In keeping with the 1950s conception of computer programming as women’s work, the two women were responsible, in particular, for all the programming — a heroic undertaking in those pre-programming language days, on the MANIAC I — and Rosenbluth in particular did all the programming for the final paper.

And also in keeping with the expectations of the time, and more depressingly, according to the Wikipedia article “After the birth of her first child, Arianna left research to focus on raising her family.”

Wall Street idealism

A Yale senior in computer science, Steph Rhee, describes an encounter with a Wall Street tycoon at a cozy Yalie networking event:

When I said that I was studying computer science because I want to be a software engineer and hope to start my own company one day, he said, “Why waste so many years learning how to code? Why not just pay someone else to build your idea?!”

What is hilarious is the imperiously aristocratic style of the grand financier, appalled at the notion of anyone getting their hands dirty in trade, in this case, being so crass as to actually develop the skills to make anything yourself, as opposed to taking advantage of your superior status to float your IDEAS into the room, and expect the peons to praise their brilliance and knuckle down to the real work.

I am reminded of Harlan Ellison’s celebrated reply to fans who would ask him “Where do you get your ideas from?”

“There’s this ‘idea service’ in Schenectady and every week like clockwork they send me a fresh six-pack of ideas for 25 bucks.” Every time I say that at a college lecture there’s always some schmuck who comes up to me and wants the address of the service.

 

A new challenge for computer science

Do you remember the Sokal affair? Mathematical physicist Alan Sokal submitted a fake paper titled “Transgressing the Boundaries: Towards a Transformative Hermeneutics of Quantum Gravity” to a postmodern cultural studies journal Social Text. Well, it was a real paper, but it was a hoax. It was written as a hoax, cobbling together the silliest tropes he could find about social construction of science and left-wing politics. Once it was published, he used it as a club to beat up on postmodern theory, and the humanities more generally. Many self-satisfied scientists claimed, at the time, that this confirmed all their worst suspicions about the emptiness of academic jargon-spinning in the humanities and social-scientists. The implication — sometimes made explicit — was that the reverse could never happen, that scientists know exactly what their terms mean, and could never be fooled by such a prank.

It turns out, the situation is much worse than that. Springer and IEEE have been forced to withdraw 120 papers that they published in various conference proceedings, and that turn out to have been randomly generated by the software SCIgen. Concerned to assure the public of the reliability of their peer-review process, Springer has now announced a technical solution: SciDetect, software that checks papers to determine whether they have been generated by SCIgen. (If it had been announced today I would have assumed that it was an April Fools prank, but the press release is from 23 March.)

Confirming me in my opinion that the whole system of peer review has outlived its usefulness, and is now living on as a vestigial parasite on the scientific enterprise.

Anyway, Springer has now thrown down the gauntlet, and young computer scientists should rise to the challenge of improving SCIgen, to fool the new software. We may see an accelerating arms race in scientific publishing, a kind of reverse Turing Test, with computers trying to fool other computers into believing that they are computer scientists. In the end, maybe the software will get so good that it will be doing original research and writing real scientific papers.

Killing the braces in order (technical)

I just noticed a funny tic that I have in programming (or writing LaTeX, which is sort of similar): When I remove brackets, I always look to remove the matched pair, even when they’re functionally equivalent, and even when it costs extra effort. For example, I have a text in LaTeX that goes something like

$\frac{1}{e^{x}}$.

I decide to change it to $e^{-x}$. So I remove the \frac{1}, and then the open brace, leaving me with $e^{-x}}$, and my cursor is right at the x. The simplest thing (measured in keystrokes) would be to shift over and remove the brace directly following the x. But there is mental resistance to removing the “wrong” brace, which I resolve by making the extra keystroke and removing the final brace. Not a big deal, but it occurs to me that there’s probably interesting work to be done (or already being done) in the psychology of programming, and the conflicts between the logical deep structure that programmers need to work with, and the surface structure that is all the computer gets to work with.

The Tory idea of education

David Cameron and his Bullingdon circle have education policies borrowed from the Wizard of Oz: Like the Scarecrow, the British public doesn’t need brains, it needs diplomas (see below). And why not? No one they know learned anything they needed to know at university except how to run away from trouble. The value of three years in Oxford for them was that they spent three years in Oxford, and that they were there together with other similarly situated scions of privileges. This is why they see nothing but prejudice in top universities’ reticence to admit the products of second-rate British comprehensive schools. They seem genuinely mystified by the notion that there could be any objective preparation that these children are lacking, preventing the top universities from admitting them to the charmed circle to which all good things flow.

Anyway, their new education initiative is to get more children learning about computer programming, dubbed the Year of Code. The director of the programme, one Lottie Dexter, explained in a recent interview, “You can pick up [learning to code] in a day.” Alas, her busy job didn’t leave her a day free to do it herself, so she knows nothing about programming, and she says she’s planning to space out this day of learning over a full year — minus the month that’s already gone. It’s not just that she lacks anything so crass as expertise in either programming or teaching; or that she couldn’t answer a question about what “code” is; or that her main qualification seems to be her excellent connections to the Conservative Party elite. Even for a non-expert her idea of what’s involved in teaching people this complicated skill are laughably vague: All about “getting people thinking about it now” and “by September they’ll be really excited”, and by the end she’s babbling about computer code as an international language fostering understanding between peoples.

Lottie Dexter explaining how to code reminded me of Monty Python explaining how to rid the world of disease.

 

 

Tech executives still lying

Marissa Meyer, CEO of Houyhnhnm? [Correction, that’s “Yahoo!”] has faced charges that her company (and other tech companies) undermined democracy and betrayed their customers’ trust by secret collusion with US espionage. She attempts to win back this trust by telling big lies. In a recent interview she claimed that

Releasing classified information is treason and you are incarcerated.

Nine words, and two (or maybe three) false statements.

First, the easy one: Treason is clearly defined in the US constitution.

Treason against the United States shall consist only in levying war against them, or in adhering to their enemies, giving them aid and comfort.

She is confusing the United States government with the 18th Century British monarchy, that classed that any release of information not authorised by the state is treason.

What her company may or may not have been constrained by is called a National Security Letter (NSL), that typically comes with a gag order. (Whether this counts as “classified information” I am not sure. I’d take her word for it if she weren’t lying about everything else.) The constitutional scope of these gag orders has been challenged in court, but I don’t know what the current status is.

But is it true that if you disclose the receipt of an NSL “you are incarcerated”? In fact, the Department of Justice — not the American Society for Feelgood Antiauthoritarianism — writes in its fact sheet on the Patriot Act Reauthorization that this act

Discourages unauthorized disclosures by providing a criminal penalty for knowing and willful violation with intent to obstruct an investigation or judicial proceeding.

Until this reauthorisation, apparently there was no specific penalty legislated for disclosing NSLs. And afterward, it’s still not clear. It would clearly — and properly — be a punishable offence if Marissa Meyer found out that her college roommate was having her Houyhnhnm? account searched because she retweeted a suspicious number of posts from Sheikh Omar’s Twitter feed, and tipped her off. But to alert the public for reasons of improving democratic accountability most likely is not illegal at all, and is the sort of calculated risk that many journalistic organisations take on a daily basis. With far less money to back them up.

I don’t doubt that some investigator fed her that line about treason and incarceration. That’s what interrogators do, they help people out of their scruples. But presumably she could afford a lawyer to give her independent advice. She could even have looked up the US Constitution and the PATRIOT act, if she knows how to use a search engine.

And I don’t expect Marissa Meyer or Sergey Brin to blow the cover off government surveillance and flee to Russia. But they clearly have decided — unlike, say, the New York Times — that they are a merely commercial organisation, with no public responsibility, and that a legal struggle would hurt their profits. That is why their customers need to make sure to align their incentives, by boycotting or simply avoiding companies that don’t show sufficient civil courage on their own.

And telling lies is not a way to rebuild trust.