On 5/20/06, Günther Noack <[EMAIL PROTECTED]> wrote:
Am 20.05.2006 um 18:33 schrieb Nicolas Roard:
> Well, additionally to the virtual filesystem, there's the whole
> notification/name service thing too... But anyway, that's why Quentin
> choose "virtual microkernel" and not microkernel :-)
> Virtual because indeed it is not an operating system dealing with
> memory allocation and scheduling; but it provides the same kind of
> services you'd expect from a microkernel..
> Anyway that's how I understand it. Interesting concept imho.
What about calling it a 'interprocess communication middleware' or
'distributed computing middleware'? CORBA does Notifications and Name
Services (and much more), too. And remember it's part of GNOME (at
least for current releases AFAIK).
The problem I got with the 'operating system' name is that it leads
to confusion about the services it is going to provide.
But.. I don't think the document says it's an operating system -- just
that we call it a "virtual microkernel". In addition to naming,
there's also the future persistence/versioning capacities of
CoreObject...
So well, of course CoreObject is not useable for the moment, so it
sounds a bit vaporware ;-) but imho the general idea behind a "virtual
microkernel" is rather interesting. And this document is here to show
the rational/idea behind étoilé, it is not a description of what we
have now but what we want to have...
So it's not an operating system is the true sense of the term, and
perhaps we can come up with a better name, why not. But "virtual
microkernel" (emphasis on virtual) is not so far off the mark, as it
is not just a naming/distributed object scheme/vfs/etc.
If you consider the sum of all we can do / want to do with CoreObject,
it's not that different from what the services of a microkernel would
provide...
I was just talking with Stefan this afternoon, and he pointed me to
keykos/eros, two microkernels that solve the persistance problem by
not doing the hard drive / ram distinction for the memory -- the only
memory you use is the hard drive, and the ram is only used as a cache.
It could be quite nifty to implement such a scheme in CoreObject, in
addition to versioning..
Anyway, it's just to say that CoreObject will be more (or could be
more) than simply a distributed object / services discovery mechanism,
and why imho the term "virtual microkernel" isn't so odd as it sounds
when you heard it first :-)
What do you think?
--
Nicolas Roard
"I love deadlines. I like the whooshing sound they make as they fly
by." -- Douglas Adams
_______________________________________________
Etoile-dev mailing list
[email protected]
https://mail.gna.org/listinfo/etoile-dev
