[<<][staapl][>>][..]
Sat Apr 12 17:13:04 EDT 2008

state.ss / 2stack

main problem: the 'data' part in state is the thing that's passed
around by all control words. this cannot be the 2nd stack: data needs
to be a stack of 'wrapped things'.. i'm not sure what that means yet,
but it's 'stuff' that gets threaded through computations.

BROOD 6 is probably going to be about doing this with composable
continuations..

i'm going to try to shield access to this data atom. i think i still
don't understand why scat-control.ss needs to have this atum clearly
visible.

trying to define these stack update functions, i find a need to make
the WHOLE state representation explicit again.

ok.. it's the right path, but i'm using the wrong abstraction.

i need a mechanism to just stick something on state to be retrieved
later. in other words, the layers of wrapping need to be made
abstract.

i.e in:    a b c D e f g H i j

if D and H interact with the threaded state in some way, they needs to
be able to do that without the lower case functions knowing about the
existence of these things.

different type of things need to work independently. or not? i.e. asm
access only makes sense if the type is actually extended with such
information.

i'm missing some crucial insight here..




[Reply][About]
[<<][staapl][>>][..]