Well my experimental program runs under Scheme w.o. crashing which means
that this is an SBCL issue. This might well persuade me to move to Scheme;
abandoning SBCL. 10 years ago I abandoned CLisp for SBCL for similar
reasons.
Scheme also compiles my set theory axioms w.o. crashing and the experiental FTP 35
was tested against FTP 34. FTP 35 has an enhanced feature for handling equivalences.
Here is the inference count on 7 problems.
FTP 34 FTP 35
1,769 233
1,899 247
545,189 9,764
2,031 13,459
1,138 26,797
16,809 852
266,721 7,439
Off to the gym.
Mark