Existencial problems on LispOS

Rodrigo Ventura yoda@isr.ist.utl.pt
Mon, 30 Mar 1998 16:58:59 +0100 (GMT+0100)

        There have been significant postings to this group about the
meaning of this project. The reason for the existence of this project
is mainly based on "dreams" people have (me included) about a
next-generation OS, based on some lisp dialect. But as someone pointed
out previously, it is extremely difficult, if not worthless, to
overcome all these year of software development based on C. The number
of people interested on LispOS is also very reduced. I guess we must
decide what we really want. I guess that the reasons why people
believe in a LispOS are mostly passionate rather than objective ones.

        To illustrate this point, let me state three trends of
discussion about what LispOS must become:

        1. LispOS is a complete OS', and so we must start by writting
a virtual machine, from scratch. This is a somehow similar approach to
Java. But if we redesign the OS based on a completely different
approach, then we'll have to re-write tons and tons of tools. We'll
have to build at least a compiler, a user interface and a text editor,
with levels of competence at least similar to Emacs. This not
difficult, this is impossible!

        2. We must start by picking up an existing lisp environment
(CMUCL, RScheme, whatever) and start re-writting well-known
tools. Someone refered porting XEmacs to scheme, and even netscape to
scheme. This also seems to me virtually impossible, if we are looking
for similar levels of competence. And how about the motivation?
Nowadays we have XEmacs for almost any platform, it's working fine,
it's an astounding piece of software. It took decades to evolve up to
this present level. It's impossible to duplicate this work in an
alternative platform, and how would you answer the question "why?".

        3. All the current OS' sucks, and the LispOS will be our
salvation. No doubt all current OS' have deficiencies, but the path to
evolution is not yet clear. We can picture in our mind some ideal OS
user interface, but the question is "how to do it". And is it clear
that a lisp based system will accomplish this task? And reiterating
the last point question: why duplicate the effort that brought current
user-interfaces to today?

        Therefore there are two perspectives under which the LispOS
project can be collocated:

        I. Make a LispOS from scratch -- the advantages of doing it
this way must ve _very_ _clear_ from the start, and until now, only
the persistent object storage is an appealing one. What are the really
great features that this approach can bring, that would be impossible
with another approach. The big problem of this approach is that we'll
have to re-write many tools that are already written in other
languages, and have been functioning for a long time now.

        II. Make LispOS ontop an existing platform -- so that we can
use already written tools, like text editors, compilers, web browsers
and a lot of other tools. The target here would be to make an
extremely sofisticated desktop, that could make LispOS "not suck",
using a popular expression in this list. But what are these features
exactly? And isn't this a desktop environment project rather than a OS

        I appologise if this post sounds mostly a pessimistic one, but
this is the way I'm viewing thing right now. It's great to discuss
ideas which seem to have little contribution to the project
development, but it's easy to get tired with it. And from a certain
point the arguments start being repetitive.

        I guess the problem is that LispOS idea is so broad, it's
difficult to focus into a specific project with a clear path.



*** Rodrigo Martins de Matos Ventura, alias <Yoda>
***  yoda@isr.ist.utl.pt, http://www.isr.ist.utl.pt/~yoda
***   Instituto de Sistemas e Robotica, Polo de Lisboa
***    Instituto Superior Tecnico, Lisboa, Portugal
***     PGP Public Key available on my homepage
*** Key fingerprint = 0C 0A 25 58 46 CF 14 99  CF 9C AF 9E 10 02 BB 2A