This damned wonderful project...

[John Casu]

The fundamental issues as I see them are as follows:

i)  Getting a working baseline prototype/development version up & running   

      soon, and deciding on what hardware platforms we want to support in   

       the medium/long term, and what lisp version/dialect we are going   
to
       use in the long term.

      IMHO, MkLinux/CMUCL/SHORE on a PC class box is the way to go to
      get something substantial working quickly; to define a direction,   
but not
      tie us down to a unix-centric view of the world too much.   Either   
MGR
      or X windows would do for a windowing system.

ii)  Being able to use LispOS as a safe delivery environment for Lisp   
based
      apps  (i.e we can reasonably ensure that it's hard to reverse   
engineer a
      vendors intellectual property), so that other people will want to   
use it.

iii)   It's got to be fast.   Equivalent canned tasks (such as reading a   
tape, or
      unpacking a tar file) should take about the same amount of time (or   
be
      faster) on LispOS as they would on a good Unix system.

iv)  It's got be be efficient.  We must be able to do serious work with   
this
      OS, on cheap hardware configs.

v)  We must have at least one killer app.  There's no point in doing this   

      project if we can't use it to do something worthwhile that's nearly   

      impossible or done badly on some other OS (even if you have a
      decent lisp environment).  Being cool is good, but not enough.

      I have no idea what this app would be :)


IHMO, the issue regarding Unix emulation/compatibility is essentially a
canard.  If we have apps and utilities to run (even if we have to write   
them),
and keystroke friendly ways  to do simple things, then that should be
sufficient to entice people sympathetic to the cause.