In the thread, Cython has no place in Leo, http://groups.google.com/group/leo-editor/browse_thread/thread/b95d8ecd38aa9717/e4662065610205eb I said: "I plan no further work on any scheme that translates Python into something else. My plan now is to put the super-lint project on the shelf for at least a week or two."
The conclusion that Cython, etc. is not for Leo is still valid, for all the reasons given in the post: Leo has no need for extra speed, and translating parts of Leo to Cython would make Leo less flexible. It's a lose-lose situation. However, I have been looking for another "juicy" project recently. You could call it "Life after Leo". The requirements for a juicy project: - It must be interesting and important to me. - It must be hard, but not intractable. - It must be open-ended, suitably vague, but well-enough defined. - It must call forth creative solutions, or better, creative problems. Imo, and my opinion is the only opinion that counts here, the analysis and (maybe) compilation of Python is a quintessentially juicy project. I love Python, and want to contribute more to it, and the over the years I have been fascinated by language-oriented questions. Furthermore, I can hardly imagine any other juicy project for which I would have a reasonable chance of making a significant contribution. The outcome of the project need not be a compiler. Some kind of program analyzer, lint-checker or other tool might arise. Indeed, inventing a new kind of collaboration between programmers and their tool might be one kind of success. I have been reviewing my posts in the leo-pylint group. My sense is that there is a lot of unjustified pessimism about what is possible in this area. The academic "proofs" that such and such is impossible are either bogus or irrelevant. This is not a computer science project: this is a computer engineering project. The fact that some initial attempts have not yielded complete solutions is also irrelevant. We do not expect to go directly from the Ford Model A to the Porsche Boxster, or from the Wright Flyer to the SR-71. A lot of work and invention is required. That's a *good* thing, in this context. Edward P.S. Yesterday I *briefly* studied the ShedSkin project. I like the FAQ: http://code.google.com/p/shedskin/wiki/faq It acknowledges all the objections to the project, but the performance numbers at http://attractivechaos.github.com/plb/ are impressive. Having said this, the fundamental ShedSkin algorithm looks like it would be hard to extend--it's more an academic exercise than proper engineering. Of course, this is only a guess, I could be completely wrong. However, my intuition is that clever algorithms are not going to get the job done. As I said in this thread, "Simple, pragmatic type inference...", http://groups.google.com/group/leo-and-pylint/browse_thread/thread/15622b4a1e11ebf I am looking for tricks rather than magic bullets. You don't go from the Model A to the Boxster with a single clever idea, you get from A to B with a lot of work. P.P.S. "Opportunity is dressed in overalls and looks like hard work." It's been variously attributed to Henry Ford and Thomas Edison. They were close friends, so perhaps they both said it :-) EKR -- You received this message because you are subscribed to the Google Groups "leo-editor" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/leo-editor?hl=en.
