OK I answered my own question with some googling.

The age of the exoplanet is not independently derived, but instead, taken from the age of the host star. This too can be difficult to determine. For isolated stars, there are precious few methods (such as gyrochronology) and they generally have large errors associated with them. Thus, instead of looking for isolated stars, astronomers searching for young exoplanets have tended to focus on clusters which can be dated more easily using the main sequence turn off method.

http://www.universetoday.com/76495/the-hunt-for-young-exoplanets/

It seems that people are deliberately creating millions of fake identities and putting them online just to screw with the bulk data collectors.

Read the explanation on this page: http://xdduk.org/nino/BT889440D

TFA states that the $460 million was lost by Knight Capital themselves. If they'd been fined $12M for stealing $460M, I'd be as outraged as the article author, but from where I'm standing it looks like the SEC turned a $460M loss into a $472M loss.

Sure, they're idiots, they've punished themselves amply!

I recently successfully persuaded the company that I freelance for to open source a core part of their product line. The part we open sourced was essentially the engine that powers several other products. I had a whole page of benefits prepared, but the main one was this:

"Your developers don't seem to realise that the core engine is supposed to be a general purpose platform, almost like an operating system - it needs to be very well documented, and it absolutely can't have any code in it that is specific to one of the applications that runs on it. If you open source it and give it its own website and code repo, your developers will finally understand what it is, and stop dumping application specific code into it when then need to implement a new application-level feature. This will save you time because you won't have to be constantly refactoring application code out of the platform."

Also, "open source is cool, and having an open source product will make it easier hiring new developers" seemed to go down well.

Google is obsessive about reducing HTML size for fast delivery, and that explains two of your observations.

The JS obfuscation is code reduction - all the variable names are replaced with a single letter and the white space stripped in all of google's JS code to reduce the script length (though no doubt they like the fact that this makes reverse engineering hard too.)

Adding the events after the page loads means you can loop over the array of links returned by document.getElementsByTagName("A"), instead of adding the handler as text to every link.