Probably not, since the ultimate goal is to have vampire find the complete proof by itself.
Currently we simply split the argument string with the function words
.
Be sure make it executable.
Use absolute paths in the scripts.
It should be possible to easily run vampire for a certain inference.
Also pass through additional options (would be helpful to add --show_active on
later).
Add option “–check-premises”. For each inference, pass the premises (without the conclusion) to vampire. If it’s unsatisfiable, print a warning that the inference is trivial (but not an error, since it’s still a valid inference). Of course it’s not decidable, but it might still help to catch some mistakes.