lispos/vm coordination

Bill House
Fri, 2 May 1997 16:49:32 -0700

Tim Pierce wrote:
>Presumably a lot of people here are following the lispvm list,
>while a lot of others (like myself) barely have the time.  Is
>anyone willing to post a periodic summary of the status of their
>project here, to improve coordination of our efforts?  For
>example, a once-a-week update might be helpful, highlighting
>decisions made, justifications, and issues yet to be resolved.
Several are reviewing a possible VM instruction set. I'm looking at Kaffe,
etc., to get a better handle on the JIT and Java issues. Also, I'm looking at
the CaptiveX Scripting API. Hopefully, by the time Mike Korns gets back from
Brazil, we'll have enough info to get an initial design spec and perhaps decide
on a code base to start with.  The goal for the instruction set (as I see it)
is to satisfy the immediate needs of the LispVM project, while allowing LispOS
to plan on provide binary compatibility for LispVM applications (whether by
embedding LispVM itself, or by translation of LVM wordcodes into whatever
LispOS prefers, nobody yet knows).

Anyway, with some good starting code and some hard work, I think we can start
to move forward pretty quickly. After all, it's not nearly so ambitious a
project as LispOS!

BTW, I really like the Flex OSKit idea as a starting point for LispOS. I'm
beginning to think that a FlexOS-based LispOS, jam-packed with connectivity
protocols and NOS services emulators, is a better approach than simply hacking
Linux. Not than I'm any kind of UNIX hacker (I'm not), but the Flex website and
arguments for a MAX-Lispy LispOS are pretty persuasive. Right now, I'm thinking
let the LispVM concentrate on infiltrating the legacy world, while LispOS
builds the proverbial bridge to the future -- kind of a "you head 'em off at
the pass while I surround 'em" strategy.

Bill House
The views expressed are mine alone,
unless you agree with me.