Thanks for your comment. I'm sorry that---despite using so many words---I didn't make myself understood. I didn't intend to argue that student evaluations can't or shouldn't be widely deployed. I was criticizing the sloppy, casual attitude towards the evidence displayed in the Atlantic article. And I thought I had linked to a literature review, and pointed out passages in the article, that supported that criticism.
Expanding the final example: say I currently give 20 students As, 15 of them score me at 1.05, 5 of them are so happy to get As they score me at 1.65. I give my 80 other students Fs, and they score me at 1.0. Then on average my A students will score me at 1.2, and my F students will score me at 1.0, and my score for the whole class will be 1.04. But suppose I were to give 10 more F students As, and they would also be so happy that they'd also score me at 1.65. By doing so, I'd increase my average score from 1.04 to 1.105---for all we've been told, that may be several standard deviations across other teachers, and may translate to substantially higher salary and job security for me.
I know more about student evaluations at the university level than at the grade levels discussed in this article. Here is an excellent overview of some of that research: http://home.sprynet.com/~owl1/sef.htm
The message one takes away from that is that (i) yes, student evaluations are a good predictor of some objective properties of a class (and other measures don't even achieve that much), but those properties aren't what teachers should be optimizing. I'd grant that (ii) it does seem worthwhile for students to see that their interests make a difference to what happens in the classroom. I'd also grant that (iii) some classroom situations may be so bad that optimizing student satisfaction may, even if not educationally ideal, still be a big improvement. And for all I know, this may be widely true at the pre-university level; but on the other hand, for all I know, giving these evaluations a big institutional role at the pre-university level could also be counter-productive...the evidence cited in the article hardly enables us to say. Any deliberation about giving student evaluations an institutionalized role should take the evidence behind (i) seriously.
One promising message from the research reported in the Atlantic article is that (iv) the specific tests being discussed have been designed in ways that seem novel and especially revealing. But the article mixed that together with an indiscriminate enthusiasm for student evaluations quite generally. And I think many people will read this and say, "Duh, that's a no brainer." Yes, it is a brainer! These kinds of policy issues aren't settleable from the armchair. Even if we cleared all the political hurdles and made someone the educational policy dictator, he or she isn't going to be able to tell just from the armchair what the results of rolling out one policy rather than another is going to be. So I get frustrated with articles like this one, that report some interesting evidence but mix it together with the kind of insensitivity to the details exhibited in comments like "That research had shown something remarkable: if you asked kids the right questions, they could identify, with uncanny accuracy, their most—and least—effective teachers. The point was so obvious, it was almost embarrassing."
Neither does this inspire confidence: "Some studies...have shown that professors inflate grades to get good reviews. So far, grades don’t seem to significantly influence responses to Ferguson’s survey: students who receive A’s rate teachers only about 10 percent higher than D students do, on average." I hope that readers of this site don't need an explanation of why the clause after the colon is only barely relevant to whether grades get inflated because that leads to better evaluations. It's almost irrelevant. In the first place, teachers needn't be aware of the cited fact; they may experience grade-inflation pressures differently. Also, the cited fact is compatible with the majority of current A-getters scoring their teachers in ways that are largely insensitive to the grades they get, but a minority of current A-getters and a majority of current B-getters being extremely responsive to the grades they get. The cited fact is just not what we need to know.
Pandoc is excellent. But it's also available for ikiwiki, since I did a quick port a year ago: <https://github.com/dubiousjim/pandoc-iki>. I don't claim it's (yet) as well-integrated as it is with gitit.
Syntax is Haselly; but capabilities are more like a mix of OCaml (reference cells, evaluation strict by default) and Scheme (dynamic typing, macros) and an algebraic language like Mathematica (expressions don't need to fully reduce).
Arch is excellent, and seems to have gained a lot of visibility/users in the past few years. And it's a great way to learn. But it is not for someone who says Ubuntu "has started releasing software a little closer to beta than I am comfortable with." Arch will look like the Ubuntu alphas. But you'll learn how to back out of and work around bad upgrades, and usually things will shake out in a few days. Doesn't sound like what you're looking for though.