Using Racket predicates like that would be a very bad idea and hard to get working as you've discovered.
I think the correct approach is to wrap the Racklog program in a Racket program that makes queries of the *log and then asks questions of the user afterwards and repeats the query if necessary. I think of *log as a database that exists at an arm's reach from the computation of import and so it shouldn't do too much work and definitely shouldn't do IO. Jay On 4/12/12, Mikko Tiihonen <mikko.tiiho...@tmtiihonen.fi> wrote: > Hi! > > I have just started to study Racket and Racklog, so this probably is a very > elementary problem, but I haven't been able to find a satisfactory solution > on my own, and I would be grateful for getting some pointers into the right > direction. > > I'm trying to employ Racklog for mixed forward and backward chaining > inference and have conducted some experimentation with it. I use a > blackboard-like repository for facts implemented as a Racklog predicate. > However I have not been able to figure out how I could automatically ask the > user for missing pieces of data. When the inference engine tries > unsuccessfully to unify a goal with the blackboard, I would like to present > the user with a dialog asking her to input the data, which could then be > added to the blackboard. Embedding this into Racklog predicates has turned > out to be harder than I could imagine - probably I've just overlooked or > misunderstood some critical piece of data in the Racklog manual. Any help > will be appreciated. > > Best regards, > > Mikko Tiihonen > > P.S. > > This succeeds, when tried directly: > > (define %pred > (%rel () > (() (%== #t (equal? (read-line) "t"))))) > > but when chained backwards from another predicate, no dialog is presented. -- Jay McCarthy <j...@cs.byu.edu> Assistant Professor / Brigham Young University http://faculty.cs.byu.edu/~jay "The glory of God is Intelligence" - D&C 93 ____________________ Racket Users list: http://lists.racket-lang.org/users
