VI4 (was: RE: [ANN]Draft rough plan for 3.6!)
Richard A. O'Keefe
ok at cs.otago.ac.nz
Tue Apr 15 23:36:00 UTC 2003
Anthony Hannan <ajh18 at cornell.edu> wrote:
Let me clarify why I think we should wait for a Jitter before
changing the image format.
I have been waiting for the Jitter for a couple of years now.
I have stopped believing that it will _ever_ happen (unless maybe
Bryce Kampjes does it).
This is a little harsh, but it's honest. It's not a case of the Jitter
always staying in alpha, it's a case of the Jitter _never_ having been
available for my machine (Sparc V8, then Ultra 1, now Ultra 2) *at all*,
and on past history, there really isn't any reason to expect that it
ever will be. Why should I put up with a 15% slowdown for the sake of
a Jitter that may never see the light of day in any form and if it does
probably not for my machine?
If I may quote http://minnow.cc.gatech.edu/squeak/3021
"What is the state of the development of J5?
It's stalled.
...
The non-optimising compiler already produces stable code
... for PPC and 386.
..."
None of this is meant to reflect adversely on Ian Piumarta.
A Jitter is a lot of work, and I haven't paid him one single cent
for any of the many things he has done for the Squeak Community.
I am not saying that he ought to have finished the thing years ago;
he has only a certain about of time and is entitled to do other
things with his life than please me. What I'm saying is that a Jitter
is a lot of work. So far it has taken a long time. Trying to target
a lot of machines at the same time would have made it even longer.
It's such a lot of work that we have no reason to expect this research
project to ship a product that I can use any time soon, if ever.
So if I can have a 15% speedup _now_, without having to wait for the
Jitter, I don't see what's so bad about providing it.
If there really is a definite plan to produce and FINISH and distribute
within say 18 months a Jitter for Squeak that works on
SPARC V9 (Solaris 2.9),
Power PC (MacOs X and/or MkLinux),
Pentium Linux,
ARM (for the sake of a major contributor to this list),
then that is *WONDERFUL* news and yes I will be happy to wait and
let's hear a lot more about it. I might even be able to find a few
days to help with the SPARC version.
More information about the Squeak-dev
mailing list
|