These algorithms all suck. Why introduce a function call penalty unnecessarily?

#define MIN(a, b) (a < b ? a : b)
#define MAX(a, b) (a < b ? b : a)

More specifically, a slow eclipse, as though there were any other kind.

Ah. So Intel is somewhat abusing the traditional meaning of a bus multiplier, and we don't actually know anything about the memory bus speed. *sigh*

However, we do know that they're running that base clock more than 10% faster than normal, which probably means that either their RAM is faster than the spec requires or they are running at a slower bus speed than the maximum. No idea which.

The other interesting thing is how this could affect the GPL. As I understand it, the difference between the GPL and say, the LGPL, is that if you write code that uses GPL libraries, your source code must also be GPL'd, even if you don't distribute those libraries with your code. (i.e. your installation instructions direct users to download and install the libraries themselves.) The way it works is, your code is written to the API specified by those libraries and you are therefore bound by the license terms of those libraries. If API's are not covered by copyright, then you wouldn't be bound by those license terms, and so effectively there's no difference between the GPL and the LGPL.

You're misunderstanding the licensing issue with GPLed libraries. The licensing issue is that by linking against the GPLed library, you are using the actual GPLed code as part of your product, not just the headers, thus making your code a derivative work of that GPLed code because it depends on that code for correct operation. Now there is some room for debating whether or not that really is the case, but that's the basic argument.

This case, by contrast, says that the headers themselves cannot be copyrighted. What this means is that you are free to do what both the BSD and GPL folks seem to do on an almost daily basis—take some library (or set of functions within a library) that is under one license and reimplement it under a less/more restrictive license, keeping the same basic interface so that code can be compiled against either version of the library and still work correctly.

I said:

Your link backs me up.

So you want to do A/B testing on whether this algorithm is better than A/B testing?

It'd probably be better to use the epsilon-greedy method when deciding whether the A/B testing or epsilon-greedy algorithm is better.

Or maybe not. Well have to test that too.

It's testing all the way down.

Neither the article nor the summary says anything about A/B focus testing. Or mention focus groups at all. It refers to A/B testing, where 2 different websites are offered to customers, and the better one found according to how objectively successful it has been. (by sales, clicks or whatever numerical measure.)

Maybe, given that most sites aren't doing it means it comes under "stuff that matters".

Don't forget the sales pitch. It could help you chose between different text. Real world trials are far better than gut feel on that.

Terminally incompetent. The slashdot website wranglers who made a system that eats up angle brackets when "Plain Old Text is selected as the format.