I didn't understand it but I found it strangely satisfying.
posted by bonobothegreat at 9:27 AM on June 4, 2009

Neat. I liked the letter "O".
posted by jabberjaw at 9:34 AM on June 4, 2009

I didn't understand it either, but it reminded me of lexipedia with wire frames :-)
posted by hortense at 9:36 AM on June 4, 2009

spooky, just last night i realized that java is a for letter word..
posted by 3mendo at 9:47 AM on June 4, 2009

Apparently they have not yet completed their analysis of capital letters.
posted by aught at 9:51 AM on June 4, 2009

Neat, but I think it ends up placing too much emphasis on alphabetical order. For instance, "dray" and "draw" are close together but "drab" is far away. But are "dray" and "draw" really connected more than "draw" and "drab"?
posted by smackfu at 9:52 AM on June 4, 2009

Cool; gave head ache.
posted by resurrexit at 10:05 AM on June 4, 2009 [1 favorite]

It says "amid" is an adjective/adverb. "Amid" is a preposition.
posted by jeremy b at 10:16 AM on June 4, 2009

Neat! I was just thinking idly about the density of permutations of letters in various four-letter words the other day, but this is about an order of magnitude more interesting as analysis and visualization goes.

The wireframe frequency dimension is a little underwhelming in execution, though—for a lot of letters the scale is such that there's no obvious information about most of the strings, since the spheres all glom together into an undifferentiated mass.
posted by cortex at 10:23 AM on June 4, 2009

If it worked on sounds instead of letters, it could be a really cool way of displaying phonotactic information. Especially if the three dimensions used were somehow spatially related to the place and manner of articulation. So...use the vertical axis for core vowel, and the other two axes for onset and coda, with glottal consonants at one end and labial ones at another. The patterns would thus be an intriguing way of representing the natural shape of English sounds.
posted by Sova at 10:31 AM on June 4, 2009 [1 favorite]

Huh. This made me wonder what the distribution of word length frequencies in the english language is. Turns out it can be modeled by a 1-displaced hyper-poisson distribution. I wondered what the hell that was, which led me to this: Mathematical aspects and modifications of Fucks' Generalized Poisson Distribution.


Circles within circles, people.


Wilhelm Fucks is now my favorite scientist name.
posted by logicpunk at 10:43 AM on June 4, 2009 [1 favorite]

Fucks' Poisson Distribution? Where's fishfucker when you need him?
posted by box at 10:49 AM on June 4, 2009

what
posted by Pronoiac at 11:02 AM on June 4, 2009

Isn't that spacial.
posted by ChuqD at 11:54 AM on June 4, 2009 [3 favorites]

You can click on the two circling arrows in the top left to stop the frame from drunkenly lurching about, which makes it a bit less seasickness-inducing.
posted by FatherDagon at 11:59 AM on June 4, 2009

Also, they make the entire thing about 'Four-letter words', and then censor "fuck"? What wuss crap. I mean, w**t w**s c**p.
posted by FatherDagon at 12:03 PM on June 4, 2009

Numerology 2.0. Which is a dismissive label I should ponder a bit more.
posted by Free word order! at 1:41 PM on June 4, 2009