Alaric B. Williams
Sun, 11 May 1997 17:25:33 +0000
> > Well, an "application" is a user-level term; the actual storage thing is
> > an independent environment, a virtual machine - dedicated to a given task.
> > IE, an "application" is a VM running a piece of code "continually" ('cept
> > it gets serialised when we don't want it).
> So we're back to the unix one-process==one-machine paradigm.
> Where applications use ipc to talk to each other, but essentially
> run in different systems.
Not quite. The applications don't run in seperate address spaces.
Just seperate namespaces. So on local machines, everything can be shared
and fast and all that. Philosophically, it's like a mutable module,
> If that's the case, then that's fine, I just thought we were talking
> about something completely different.
There's "we" and there's "me" - I'm just outlining one model of
dividing processes up into administrative units.
How about treating the "applications" as modules, then, and using
::-like syntax? You lose on simple network transparency, ok, but
these are tradeoffs, and somebody might have a model that allows
both without some kludge like distributed shared memory! (Any offers?)
Alaric B. Williams (firstname.lastname@example.org)
---<## OpenDOS FAQ ##>---
Plain HTML: http://www.delorie.com/opendos/faq/
Fancy HTML: http://www.deltasoft.com/faq0000.html