Tags: unusual words

Dwight

Look at me invent new words!

From my log file:

There is a flaw in this argument. If there is a core that sustains memory, then some quale might have purpose -- not as the quale itself, but as the results it brings to the core. Enlightenment is a good example of this, although it's not directly qualic.

At least they're extrapolations of common suffixes; nothing as horribly twisted as [that trendy word for a web journal] or [to search on the web, a verb of a company's name].

And on a similar note, I think I've experienced eku overload. It's kinda like someone in a desert happening upon a random airdrop of a few thousand barrels of water, only to become personally acquainted with water intoxication.
Dwight

And while my screen is filled with hexadecimal numbers

I got the idea that any tree search can be turned into a game. The concept is simple: you have a "seeker" and an "obscurer". The seeker tries to find the right answer by the rules governing the search, by asking the obscurer. The obscurer tries to adversarily pick a response that ensures the search will take as long a time as possible - within the constraint that he cannot contradict himself.

The game version of binary search is the old "guess a number and I'll tell you if it's higher or lower than what I'm thinking". The consistency constraint means the obscurer must in the end reveal a number that matches what the seeker asked: for instance, if the seeker asked if it was greater or lesser than 100 and got the answer that it was greater, the number must indeed be greater. But the adversary nature means that at the first step, the only thing the obscurer commits himself to by saying the number is greater than 100 is that the number is greater than 100.

For binary search within a range, it's rather easy for the obscurer to answer in a way that means the seeker has to ask all log_2(n) questions even in the best case, so the optimal strategy is simple for both. But for more complex searches, the strategy gets more tricky too. Consider hangman; at the first step, the seeker asks "are there any Es in the word, and if so, where are they?". The obscurer can now say that there is no Es, but then he has limited himself to the words that contain no Es. It may be better for the obscurer (for a long term strategy) to pick the space of words with only one E, or three Es, or whatever, but what space is best depends on how the seeker will ask next. A classical "guess my response to your response to my response" game.

(Even if the obscurer knows the seeker's heuristic, as in minmax, these guess-responses games can have complex strategies.)