The update on how the performance characteristics of this thing changed after JIT would have been interesting. The paper’s 11 years old now.
There’s a follow-up paper (2010) with the JIT implemented and reporting some benchmarks. I don’t think the project continued after that, although the lead author has continued research in related areas.
Oh darn, I didn’t realize he was also an author of a paper I was saving for next week when more readers are on. He’s doing some clever stuff. Interestingly, he also ported Squeak to Python while Squeak people made it easier to debug Python using Squeak’s tooling. Some odd, interesting connections happening there.