search

Lexpedite / blawx

A user-friendly web-based tool for Rules as Code.
MIT License
105 stars 9 forks source link

Fix how interview endpoint deals with defeasibility #315

Closed Gauntlet173 closed 2 years ago

Gauntlet173 commented 2 years ago

Right now, the interview endpoint is running a no-abducibility version of the query in order to determine if there are any correct answers, and returning them if there are. However, that doesn't work with the defeasibility system.

Need to change the way the interview endpoint determines what constitutes a solid answer. Might need to check for answers without abduced statements, or potentially for the existence of opposite models?

Gauntlet173 commented 2 years ago

This problem seems to have been solved by using only the query with abducibility in the interview reasoner, instead of excluding abducibility for finding actual answers. Now if there is an exception, the example is only contingently valid until that exception is eliminated or no longer abducible.