> > a) this might require omega to be more stable than omega users would
 > > wish, ie it might prematurely limit addition of new features.
 > I can't see why it should limit additions, as long as old things work as
 > they worked.

in the context of something like LaTeX (which defines a Language just like C)
additions means compromising the exchangeability of documents between
different implementations. I could bet that even with the few peoples on this
list we will find more than 3 different releases of LaTeX being the production
one at their site/computer. Nevertheless we expect that essentially any LaTeX
document those source is being send to the list would be processable by
everybody (and unless you trying deliberately to build on fixed bugs etc it
will be true). That is not at all true if you keep adding features which will
get added into the kernel or package macros so that only those people with the
latest release will be able to process stuff.