Here are all of the confirmed sources for information about the Prizm:

Casio general specifications

Hardware User's guide

Casio Software user's guide

Spansion Flash memory Datasheet

Pictures for graphing (proprietary format)
Also, all of the video animations for the Prizm released by Casio have been sourced.

Bar chart source URL: Well, a graph of some sorts was to be expected.
Size: 189 Kb
http://us.fotolia.com/id/19974173

Hot Chocolate stream source URL: Pretty neat. I'll be impressed if the Prizm can properly display this.
Size: 112 Kb
http://us.fotolia.com/id/23691097

Clock Time Lapse: If watching clocks is your thing...
Size: 443 Kb
http://us.fotolia.com/id/13760669

Coaster car: I don't understand this one either.
Size: 927 Kb
http://us.fotolia.com/id/18026842

Hearts: There isn't really much I can say about this one.
Size: 662 Kb
http://us.fotolia.com/id/21628036

Jumping Fish: Kind of cool, actually. I like the backflip.
Size: 285 Kb
http://us.fotolia.com/id/13162206

Pendulum: WTF?
Size: 816 Kb
http://us.fotolia.com/id/13446446

Wind Turbine: Cool shot, but I'm not sure I would waste the memory on it.
Size: 581 kb
http://us.fotolia.com/id/17327954

If you are aware of anything not in this list, please post a link to it.

Additionally, it is strongly suspected that the Prizm is Big Endian, or at least uses Big Endian formatting. Furthermore, the .g3p file format is technically broken. These are both unofficial until tested and confirmed, though.

This is awesome information. Do you have any docs on the broken format?

Yes, but all I have is the Hex. Plus, it's only the security sequences/header that have been cracked. I'm still working on replicating images.

Thanks so much for sharing this! I notice that the serial port is listed as a "three-pin" serial port - I wonder if we could bit-bang a connection over something like CALCnet.

Is three-pin the kind of serial port 8x uses as well?

If by "three-pin" they mean "two pins and a ground", then it's almost definitely bit-bangable in ASM into something TI-compatible. If they mean "three pins and a ground", then no such luck.

You couldn't just ignore a pin?

We can make an adapter if it's the latter. But wouldn't it be incompatible with most computers if it were the latter?

Oh, good point. (/me facepalms) The thing you couldn't do is try to bitbang the TI protocol to keep up with the Casio serial protocol. I think it is a 4-wire, 1 ground protocol, because they mention serial with hardware flow control, aka TX/RX; RX/TX; XON/XOFF.

This isn't really documentation, but it's not worth its own topic:
WFRNG was ported to the Prizm a little while ago.

http://ourl.ca/8029[/url]

I'll be very honest with you; I'm somewhat sick of the WFRNG meme, but I'd glad to see that people are learning this new device. What language was the WFRNG clone written in?

BASIC. Casio hasn't released the C compiler yet.

Gotcha, that makes sense. How does the Casio flavor of BASC compare to the TI-BASIC variant, if that's not too stupid of a question that I could answer myself with a bit of JFGI.

In terms of functions, it's [almost] everything TI-BASIC should have been. The only things missing are the USB port and spriting commands.

Oh wow, that's great to hear! I'm thinking more and more strongly about getting one of these when they come out to play with; how many of you are doing the same?

If I have any money left after Christmas I might get it

Might not be a bad idea, but I'll wait for you guys to figure out the hard stuff so I can do it easier.
I know, I'm lazy.
(Not really.)

Well, there's a difference between (1) being lazy, (2) not knowing enough to do something, and (3) being too lazy to experiment and learn.

Well, that's why I'm doing it. No need for everyone to learn how to edit long hex codes to make their 1337 BASIC quadratic solvers [with large graphics!]

Oooh, color quadratic solver! Can't wait for the first one of those. >_<