Is there something clear or latent method, procedure, SUBL
command(s) or something else for debugging or viewing _full_ inference tree
generated by OpenCyc inference engine?
What do I mean?
Imagine that somebody starts some query to Cyc like
(CYC-QUERY '(#$isa #$Animal ?SMTH) #$EverythingPSC)
And receive some number of bundles like ((?SMTH . #$Collection)).
So, OpenCyc internally construct some inference tree for this prove.
This inference tree consists of proving nodes (goals, dead-ens,
intermediates), used rules, gafs etc - all are used during proving process.
And this inference tree, as we know, contained in some data structure inside OpenCyc
along with some bookkeeping information.
Standard "Ask/Query Tool" which you can find in OpenCyc web interface lists
only short bit of such information: only goals (not dead-ends) and rules
and gafs that lead for each goal.
But I prefer more. It's desirable to receive full or almost full information about
inference tree (all success and unsuccess proving attempts, reasons of
conclude something on every prove step and so on).
So, I want some feature which can, for example, logs all information
during inference process into the some log file (e.g. cyc-log.txt). By other
words, full (or almost full) debugging information about OpenCyc inference.
Unfortunately such feature is not stated clearly in web-interface, SubL
API or any OpenCyc documentation that I have read.
Of course, somebody can cracks (brute force) OpenCyc, learns with
proceedings which are taking place in computer memory during and after inference
process and after tons of hacker work lists a full inference tree, may be.
But I think it isn't a good idea.
I think in "Big" Cyc such feature exactly is from the beginnings of Cyc
project. Because which other way can cyclists analyze wrong queries
and assertions? Right! Only print all inference trace (presented in
machine-generated view) in text file or something like this! May be
such feature doesn't accessible in OpenCyc version of Cyc?
Of course, it so hard and laborious to analyze manually full inference tree
instead of reading "Ask/Query Tool" output. But I need this for
my own research project connected with automatically analyze absent
information in OpenCyc KB and putting it into OpenCyc also
If I will decide pointed automation problem I will present results of
my research for all interested peoples.
Thanks for any comments!
Moscow State University of Economics, Statistics and Computer Sciences.