subspace
phext makes use of an 11-dimensional text space that i affectionately call subspace (as a thank you to star trek for expanding my mind as a child). if phext stopped at subspace, it would work quite well for large language models, but not much else. in order for humans to make sense of subspace, we need a mechanism for walking phext spaces.

consider for a moment how 2D text is presented on a screen. as you type characters from subspace, they are rendered in positional order. there are no indexes or offsets to consider - everyone agrees that each byte advances your position by 1 position to the right*.

the humble line break introduced a major design split between the pc industry and mainframes. in the 1970s, computers were very limited in available memory. a computer with 2 KB of memory versus 4 KB of memory might double the cost of your purchase! note that a single page of text is usually 80 columns x 25 characters - about 2 KB.

it turns out that most lines are empty space, though. so unless you're writing war & peace, you can probably get by with 1 KB of memory. this cuts the cost of your early computer by quite a bit!

ok, so back to today: we have terabytes of disk space and gigabytes of memory - why does this matter?