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
This explanation will probably be hard to follow, as it relates to implementation details that are not well documented yet.
The current implementation of tableau->sequent conversion has the following constraint: A tableau which refers to an earlier row more than once on the same branch will cause the conversion to fail.
This is because, at each node, the conversion removes the cited formula from the array of formulas which is passed down to the next recursive call, and because the row cited by a node must appear as the row of one of the formulas in the current array.
Changing this probably requires changing the sequent rules on the Carnap side so that the formulas being resolved must be repeated in the sequent premise. However, this would cause a problem with the 'Lit' rule.
It may be necessary to have different conversion behaviour for the SL and QL checkers.
The text was updated successfully, but these errors were encountered:
Reopening since this can be fixed prior to actually having a checker to support #87.
Having thought about this more, the way to do it is probably to keep track of all the formulas, including previously discharged ones. Then we can use 'Struct' or an explicit weakening rule to add them at the stage where we reorder the formulas.
This explanation will probably be hard to follow, as it relates to implementation details that are not well documented yet.
The current implementation of tableau->sequent conversion has the following constraint: A tableau which refers to an earlier row more than once on the same branch will cause the conversion to fail.
This is because, at each node, the conversion removes the cited formula from the array of formulas which is passed down to the next recursive call, and because the row cited by a node must appear as the row of one of the formulas in the current array.
Changing this probably requires changing the sequent rules on the Carnap side so that the formulas being resolved must be repeated in the sequent premise. However, this would cause a problem with the 'Lit' rule.
It may be necessary to have different conversion behaviour for the SL and QL checkers.
The text was updated successfully, but these errors were encountered: