Why [not] X?

Jordan Henderson jordan@Starbase.NeoSoft.COM
Thu, 22 May 1997 22:06:47 -0500 (CDT)

> From: Mike McDonald <mikemac@titian.engr.sgi.com>
>   Actually, his initials are, *ahem*, MRM. :-) I just think you guys
> have dived off the high board head first with Henry and Fare judging
> the dives. Only problem is no one bothered to check to see if there
> was any water in the pool!
>   Mike McDonald
>   mikemac@engr.sgi.com

Now, THIS, I agree with.  

I may be guilty of it myself sometimes (see my recent ranting about
the superiority of NeWS to X), but this list is really just becoming
a big waste of bandwidth.  As Patrick Logan has said "I'm just trying
to figure out what this thing you are trying to build >IS<."

Nobody has a direction or a goal in mind here.  So far I think we want a
POS-based, completely reflective, minimally dependent on Unix/C, 
high-performance, completely portable, reimplementation of LispMs,
OODB integrating, immutable cons-box sporting, scheme & CL & Eulisp & Dylan,
supporting all known high-level languages as programmatic transformations
to a "basic Lisp", thread supporting, network transparent, web browsing, 
no file system as we know it, Artificially Intelligent, Lisp OS/Environment 
with it's own VM.  Sorry if I left anyone's strongly favored features out.

I think we can begin with the end in mind, but that doesn't mean that we
start on the end product today.  We could learn a lot from the FSF.  They
stated their intention of having a free Unix clone from the start, but 
what did they focus on?  INFRASTRUCTURE.  We need to develop a set of
kickass, highly portable Lisp environments and applications that are 
largely freely available to increase Lisp mindshare.  Our ultimate goals
(whatever those might be) cannot be fulfilled by the people on this list.

If we had better tools in hand we could START to prototype in a 
few directions toward the ultimate goal.  One direction would be just to
get CMU-CL up on the Flux toolkit, so that we have a firm idea of what
is lacking and what is needed in Lisp languages to support low-level
OS programming.  Another direction would be to support the cons.org efforts
to port CMU-CL far and wide (WindowsNT/95 anyone?), so that more and more
people can be using Lisp and see it as an efficient, compiled language.
Also, we could move TOWARD hosting everything on a LispVM.  The LispVM
could be initially a plug-in for today's web-browsers, but tomorrow 
maybe it could serve as a portable assembly language for implementing
the whole tool chain/OS.

There are more good application ideas here than there are people to work
on them.  I believe that if we had a set of tools we could get investment 
in time and money from many sources to continue to develop this work.  

Discussion is not a bad thing, but much of what we are doing is just 
oneupsmanship.  People often fall into the trap that the only kind of
intelligence is critical intelligence, the ability to criticize someone
else's ideas is seen as confirmation of worth.

I still think my proposal archive is a good idea.  If people have specific
proposals for some implementation issue/feature/application, send them to me
and I'll put them up at http://www.neosoft.com/~jordan/LispersAnonymous.  
Then, people can read the proposals and send their comments to the author of
the proposal.  I would hope that this would lead to Birds-of-a-feather 
groups forming and teams going.  But, I don't know.  I'm beginning to think
that nothing will come of all this talk, just as no palpable progress is
being made on similar projects you can find out there.

Don't mind me, I haven't had my medication today.