NIkita, did you...

check out Sussman's new programming paradigm? Might give you some nice ideas for Symta before it reaches a stable release.

http://web.mit.edu/~axch/www/ilc09-slides.pdf

http://groups.csail.mit.edu/mac/users/gjs/propagators/

https://github.com/ProjectMAC/propagators

http://dspace.mit.edu/bitstream/handle/1721.1/44215/MIT-CSAIL-TR-2009-002.pdf

>>19
They are useless for general purpose programming, because every time you change root value, you have to propagate change to all dependent value. This is even more inefficient than Haskell's thunks.