Luis M. González wrote: > I'd like to know, in your opinion, how far is the goal of making pypy > complete and fast?
Me too :-) PyPy is doing a great job, that's for sure. I'm hesitant with making estimates, after I learned what a bad job I'm doing at extrapolation. First I thought that we would reach our first self-contained PyPy much earlier, gaining CPython speed. When I had lost my faith a little, we suddenly made it. Then we made very much progress in speeding it up, but still we are 7 to 10 times slower than CPython, and it gets harder and harder. Now we are aiming at JIT technology, which is able to accelerate Python quite much in many cases, even if we should fail to improve the basic translation reasonably. Of course it would be nice to reach both aims, and I expect that the things we will learn from writing the JIT will also improve the static translation. Completenes? In some aspects, like CPython compatibility, we are very complete, maybe more than the original, even. :-) Concerning the promises we made to the EU, we will have a hard time to make it all happen on schedule, but we have a chance, given that the support by external helpers keeps growing. Concerning all what we ever said about PyPy? This is a never- ending story and unlimited, as I don't expect PyPy to stop growing and extending in any near future, like Python doesn't... > Regarding the current state of the project, are you confident that the > goals will be met, or you still have doubts? I no longer have doubts about success. I never really had, but my time estimates are less pessimistic as they sometimes were. I don't really believe that we will outperform CPython with a translated RPython interpreter by the end of next year. We will probably, in conjunction with a JIT compiler. For gaining a maximum of performance, my guess is another two years would make very much sense. > (The same questions go for the RPython translator project as > stand-alone tool)... This is a matter of viewpoint. As a developer, I'm able to create extenson modules on demand without any explicit tools. Enabling/supporting the most needed features might be doable in a couple of weeks and months, depending on the expectations. A simple-to-use, stand-alone tool for making extensions will maybe not happen at all, unless we get a lot of extra-resources. I'm expecting something to happen in the first quarter of the year. It depends on how much we can extend activities without missing the promised goals which we have to fulfill, and how much sponsoring we can create. I believe that by providing just enough support to make some companies productive in using PyPy, we will create enough funding for the time after 2006 to make PyPy survive for a long time, and creating tools like this will become a self-running motor for PyPy. A matter of good balancing :-) merry christmas -- chris -- Christian Tismer :^) <mailto:[EMAIL PROTECTED]> tismerysoft GmbH : Have a break! Take a ride on Python's Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 whom do you want to sponsor today? http://www.stackless.com/ -- http://mail.python.org/mailman/listinfo/python-list
