Dear Jan,
Congrats for the great work on s(CASP).
However I see that comparing SWI vs Ciao in the performance table is a bit misleading, since it is not SWI vs Ciao but just comparing two versions of sCASP?
Good remark, Jose. Sorry. I’ve added a disclaimer to the original text.
The naive initial port of s(CASP) was faster on Ciao than on SWI-Prolog. I do not recall the exact difference. I think somewhere around 30%. That is probably not fair either as in part that was caused by the port using Prolog emulation for Ciao built-ins. Prolog systems are pretty hard to compare on performance as there are many dimensions.
Thanks!! It is awesome that one can edit posts 
I wonder whether the “turbo” data could be expanded by a column that shows execution speed with erased justifications. So that only the dual logic and abduction performance is measured? Does s(CASP) have such an option?
Also I am taking notes about an idea for justification on demand that just appeared: What about using clause ref for this purpose. The missing link so to speak, the synthetic argument that would always help to make a justification drill down?