In the Oz world, the method of solution is programmable through the concept of Computation Spaces. Oz has the freedom to program different solvers that interleave the computations in any number of ways. As I wrote elsewhere, "the actual strategy in Oz is determined by the Solve function (which corresponds to the "run" function in Scheme), not in the declaration of choice (conde, condi, conda, condu)". It should be quite feasible to program a solver in Oz which corresponds to the various interleave choices provided by Kanren.

On the translation front, I haven't got much further, mostly owing to the time constraints of working on several startup ventures. Where I left off was the point where I needed to figure out how to program search strategies beyond the out-of-the-box depth-first search Solve function. Russ Abbott did provide me with a breadth-first solution:

http://cs.calstatela.edu/wiki/index.php/Courses/CS_460/Fall_2005/SolveAll

As I recall his solver did correctly do a breadth-first search, but it didn't return the solution list in the same order as condi. I started to read Christian's book but it was about that time that I got off-track and haven't seen my way clear to complete the project yet.

In the meantime, perhaps others that are more familiar with Oz can provide a better overview of how Oz uses Computation Spaces to achieve various interleaving strategies.

Thanks,
Chris Rathman







Daniel Friedman wrote:
Chris,

I was wondering if you tackled any of the arithmetic code of chapters 7 and 8. The reason I am curious is that we use interleaving computations to get that to work and I don't know how OZ handles that. It would be very
educational to see how that works in OZ.

... Dan

_________________________________________________________________________________ mozart-users mailing list [email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users



_________________________________________________________________________________
mozart-users mailing list                               
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users

Reply via email to