Linas Vepštas

arghhh ``` (define (gaussian m d) (NumberNode (+ (string->number (cog-name m)) (* (string->number (cog-name d)) (random:normal))))) ``` I actually tested that, and it works.

I assume that "hangs" == 0% cpu use, and not 100% cpu use ... I assume that stack traces show a bunch of threads in deadlock. Does it hang if...

The idea of the atomspace is that it should be impossible to make it hang, no matter how buggy r2l-parse or the forward chainer is. ...

FWIW, at this time, sureal is unmaintained.

Hi Arley, ... and Nil, I think we need to fix contin_t for "ordinary" moses, not just particle-swarm moses. How about this idea: we store a full-length float or double...

On Tue, Jul 7, 2015 at 8:17 PM, ngeiswei [email protected] wrote: > @linas https://github.com/linas the current state of the code already > nearly does that, > > Arghh... yes but...

p.s. what you call a "contin trit" is called a "Cantor tree" in the general world. http://blogs.scientificamerican.com/critical-opalescence/files/2012/08/f5.jpg The cantor tree unfolds into the "modular group" and there is all sorts...

Hi Nil, Arley, I have to retract/modify my earlier comments as well; they were based on a misunderstanding of sorts. So, what I meant to say was this: -- I...

OK, sure. My knee-jerk reaction is "simplify the code" and remove contin entirely, but I'm tired just right now and don't really care. :-) Do the right thing.

What is the test problem that you are running? --linas On Sat, Jul 11, 2015 at 9:53 AM, arleyristar [email protected] wrote: > TL;DR update: > Problem, worse performance, don't expect...