🚀 go-pugleaf

>From Craig.Everhart@CMU-10A Sun Nov  1 23:01:49 1981
You're suggesting that it's the smallness of the physical address space that's
a problem.  Only operationally, I reply.  The programming problem is the small
virtual address space, for holding program and data both.  It's good to avoid
having to interpret every reference, where the memory address is large and
the hardware-provided addresses are themselves small.

This was one of the principal problems with programming Hydra; it's isomorphic
to the 11-90 problem RMS mentions.  It's unfortunate for the programmer to have
to worry about interpretation of wide addresses at each step of the program.
Interpreters that provide a wide-addressed virtual machine are cute but slow.