Putting salt on food before you taste it

There's a story about a woman who went to a job interview, and was taken out to dinner, and put salt on her food before tasting it, and the boss immediately rejected her without further consideration.

It didn't literally happen, but the attitude that you should or shouldn't add salt to food is something people really do argue about.

The intended message is something like, "the candidate was stupid for adding salt without knowing if the food was salty enough, since she can't take it out again".

But in fact, that's an etiquette issue. Do you trust the person preparing the meal to have prepared it correctly, or not?

But as with all parables, there are many possible messages. What I see is a class issue. If you have been raised from birth eating primarily meals cooked by a personal chef, then yes, any meal in front of you should be tailored to your personal taste, and you shouldn't assume otherwise.

But if you've been cooked for primarily by an overworked unaided parent trying to cook for seven, or by McDonalds, the food probably hasn't been tailored to your personal taste. Or it has in some respects, but the chef probably aimed for "slightly below average preference for salt"[1] and assumed that everyone would know that and could add salt to whatever level they preferred.

There are reasons for workplaces to require people to jump through hoops even if they're completely arbitrary (eg. wearing shoes, wearing ties, not swearing, lying and pretending personality tests are super-effective, using a formal register of speech, etc). At a minimum it selects people who are willing to put in effort to fit in and not be disruptive.

But it's also true that, as a side effect, it selects for the people who are already in that system, and against those who aren't, even if they're equally competent.

So I'm not sure if I blame the fictional boss for blaming the applicant. I'm not even sure if I blame her for lying and claiming the applicant stupid, rather than just for coming from a different culture -- claiming that the ways people signal high social status are due to inherent virtue rather than learned conformity is itself a way people effectively signal high social status!

But even if it's inevitable that people do so, and even if I'm equally "guilty" of being "stupid" and of "eating in the wrong restaurants", I resent being dismissed for the wrong thing, for being called stupid if I come from the wrong culture, or having my culture blamed if I ever show mental aptitude for something.

[1] In taste, not health.

Open-access maths journals

http://gowers.wordpress.com/2013/01/14/why-ive-joined-the-bad-guys/
http://gowers.wordpress.com/2013/01/16/why-ive-also-joined-the-good-guys/

Prof Tim Gowers recently posted about two new open-access maths journals he's involved with.

History

There was recently an ongoing fuss about academic journals. The general sentiment seemed to be, that academics did a lot of the significant work in preparing journals, in refereeing papers for technical accuracy, sitting on editorial boards, etc, and the remainder, of copyediting, formatting, physically printing and distributing, while valuable, wasn't worth as much as it had come to cost.

Fingers were pointed especially at Elsevier, who many people claim to represent the end of the publishing industry who have bought the most small journals, and squeezed the most money for them with the least commitment to actually supporting the acadmic community. Especially by charging inflated prices for necessary journals, spuriously justified by bundling them with a lot of other less-necessary journals. And by forbidding academic libraries from discussing what they have to pay to subscribe to the journals, which isn't inherently bad, but is generally a pretty clear indication that you're doing something dodgy.

In fact, comparisons to smallpox were thrown around. Even if Elsevier's capitalism wasn't evil, it was stupid, in that if you kill of the host, your parasitic infection is inherently doomed. I don't know if this was justified (but I assume that Tim Gowers was right).

Journals are not funded by people who buy individual issues. They're funded by university libraries who subscribe. And universities all do have to subscribe, since you can't do much research without reading other people's research. So lots of people said, "since almost everyone pays for them and almost everyone can read them, why can't we cut out the middle men, fund the journals directly without the fiction of subscribing, and just throw them open to everyone?"

The trouble is, what do the existing journals have? They have a reputation. Academics all know that if you're published in Nature, that's good. Going begging asking for money to set up a new journal less famous than nature doesn't get very far, if everyone ALSO still has to pay subscription fees to the existing journals.

The only way to break the deadlock is to make a sufficiently high-profile break with the previous system that at least one new journal instantly acquires enough reputation to be unignorable. Which is why people have been making a fuss about it.

I only know about maths, because I read Tim Gowers' blog. Maths is a bit of a weird edge case because mathematicians have to pay pretty much only for journals: otherwise, you just need an office, a pen, a wastepaper basket, and some grad students, and those generally turn up in universities.

Forum of Mathematics

There is a new author-pays open-access family of mathematics journals set up by Cambridge University Press. The idea is, instead of being charged to read a journal, you're charged to publish in a journal (since both are equally necessary to an academic career). And anyone at all can just read it.

And that institutions will set aside a pot of money to pay for the publication, just as they previously set aside a pot of money for journal subscriptions. I think some other subjects already work like this?

The controversy is, Tim Gowers says the journal will continue to accept anyone's submissions if they're good enough, and won't be rejected if the applicant can't pay. And that any sufficiently prestigious university will just automatically pay the fees.

Which sounds reasonable to me -- presumably most publications come from big universities, which will automatically do the done thing.

However, detractors assume any author-pays model will be like more vanity publishing, and automatically deter authors who can't pay and don't work for a big university, or people from smaller or less rich institutions.

The comments section

Prof Gowers' post, and the detracting blog post he linked to are obviously sensible. I sort of assume Gowers is right, but I don't actually know.

But the comments section is so depressing. People going out of their way to comment on a the blog of a high-profile research mathematician are better formatted, but just as puerile, as those of many other blogs. Everyone talks past each other, saying, author-pays WILL AUTOMATICALLY be vanity publishing, or author-pays will just be business as usual for large institutions, and assume the other is being wilfully evil by deliberately ignoring the obvious truth, while never actually proposing any evidence for what it will actually be like.

Next time

Next time: the ArXiv-only journal.

Open-access maths journals #2

An ArXiv-only journal

The process for publishing an acadmic paper goes something like:

1. Do lots of hard thinking no-one in the history of the world has ever done before
2. Write it up neatly
3. Submit it to a journal, which will have someone check it's (a) new and (b) interesting
4. Wait a long, long time
5. Profit.

OK, I was lying about step 5. The point is, mathematicians, possibly even more nerdy than most academics, having got through step 1 and some of step 2, sort of get fed up at the point where they have to talk to someone and don't want to bother with it any more.

A sort-of short-circuit in the process is arXiv, an online repository where, when your paper is going to be published, you can post a preliminary copy, so people can see roughly what's going to be in it without having to wait for the journal to be actually printed.

At this point, a cynic might start asking, do we actually have to do steps 3-4?

In fact, one of the recent big breakthroughs in mathematics, Grigori Perelman's outline proof of the Poincare conjecture, one of the Millennium problems, for which he was awarded (and rejected) a Fields Medal, was published entirely on arXiv. There's some controversy whether the "proof" owes to Perelman's outline, to the people who published a complete proof, or some combination, but everyone accepts that Perelman's contribution was very important.

However, Perelman's proof was clearly an important and serious effort. For other papers, there's still a value in peer-reviewing, editing, sorting, collating, etc.

http://gowers.wordpress.com/2013/01/16/why-ive-also-joined-the-good-guys/

Gowers is also involved with a completely free journal which has done what many people wonder about, doing the bit of collating unpublished papers into a "journal", but leaving off all the other bits, so you just get a list of "these papers on arXiv are true and useful," but they're just as available to read as anything else on arXiv.

It remains to be seen how successful this will be. We'll probably end up with some compromise with different sorts of journal for different purposes, but hopefully a further improvement over the current system. But its good to see a fairly high-profile attempt.

I think PLOS in biology is something similar, but I don't know how similar?

Bridge: opening hands

I did this a little before, but I thought I'd try to set down a little of the system I usually end up playing, which is a strange hybrid of what-CU-bridge-club-plays and what-ex-ncipher-people-play, leaning one way or the other depending on who I'm playing with.

Not as a definitive guide, but at least, as an attempt at introspection at what I do do.

Targets

Evaluate your hand with high card points and loser count. For a 3NT game, you need approx. 25 high card points between you and your partner. For a 4H/4S game, you need to have a trump fit, and then 14 losers max between the two hands.

For a 5C/5D game, you need 13 losers max, which is quite hard to find -- in practice you often need to be even stronger than that to avoid losing three tricks. Because I play a lot of duplicate bridge, I almost never actually try to bid a minor suit game unless (a) 3NT is definitely a bad idea and (b) we clearly only have 13 or 12 losers. However, it's a bad habit to never look for minor suit games, if it's the right contract, ignore me and bid it.

When to open

( Read more... )