You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, there is a class Result that is returned by the evaluate function of a Constraint, and there is a related class Evaluation that contains (I think) all the information in a Result, plus additional information.
This seems redundant. We should just have evaluate return an Evaluation, with most of the fields not yet specified, and the search algorithm can later populate those fields rather than making a whole new object.
I think the reason they are separate is that I used to make an Evaluation only for constraint violations, whereas a Result was created for every evaluation of a constraint. However, that's no longer the case. We keep a list of all Evaluation's. But there still might be a good reason to keep these as separate classes.
The text was updated successfully, but these errors were encountered:
Currently, there is a class
Result
that is returned by theevaluate
function of aConstraint
, and there is a related classEvaluation
that contains (I think) all the information in aResult
, plus additional information.This seems redundant. We should just have
evaluate
return anEvaluation
, with most of the fields not yet specified, and the search algorithm can later populate those fields rather than making a whole new object.I think the reason they are separate is that I used to make an
Evaluation
only for constraint violations, whereas aResult
was created for every evaluation of a constraint. However, that's no longer the case. We keep a list of allEvaluation
's. But there still might be a good reason to keep these as separate classes.The text was updated successfully, but these errors were encountered: