[MUD-Dev] Re: DevMUD Prototyping

[Niklas Elmqvist]

Note:
I expect we'll move this discussion to private e-mail after this, but I
thought we might discuss this initiative on the list so that people are
aware of it...

[Jon Leonard:]
> >and that we're not
> >likely to see alternate proofs of concept in the near future.
> >(Even though I would welcome seeing one.)

[James Wilson:]
> well... I would really like to see someone (Nick?) try to prove
> me wrong viz. Gamora's efficiency. CORBA would also be interesting to
> benchmark. (Nick, would you be interested in throwing something
> together with me so we can see exactly what sort of costs are involved?
> I am really quite curious about this.)

Really? *blinks tears out of his eyes* Do I get to do it message-based
with a Gamora-like architecture? :) 

In all seriousness, yes, I am definitely with you on this, James. It might
be useful to have something to point at when we want to compare the
C-based DevCore with an OO-possibly-message-based C++ DevCore. If nothing
else, maybe some of our ideas can make it into the final DevCore, and some
things might be converted for use in DevMUD modules. I, for one, think
that if a message-based communication protocol doesn't make it into the
core, I will probably write a module to do it. 

I do happen to have some code (the beginnings of an event manager and a
base event class) in this vein which we could take a look at (remnants of
my original MoleMUD and MoleCore project). Not much and not necessarily
good, but I actually toyed with the idea of trying to make a prototype out
of it earlier, but the list's arguments about the disadvantages of a C++
message-based architecture changed my mind. 

We should probably aim at providing about the same minimalist feature set
as Jon's prototype so that we don't waste a lot of time adding
functionality we very well might end up discarding. It might also be
useful to implement some kind of testbed to evaluate performance -- that
is, check how many messages we can pass around without bringing down the
system to its knees. Maybe there is a way that Jon Leonard's prototype
could be tested in the same way to address uncertainities (especially the
ones Ola has pointed out) about the performance and optimization of a
dynamic-library-based system as well?

Oh, and using CORBA sounds *very* interesting, but I have absolutely no
experience in the field, so you'll have to take the lead there. 

> James

-- Niklas Elmqvist (d97elm at dtek.chalmers.se) ----------------------
  "The trouble with being a god is that you've got no one to 
   pray to."
		-- Terry Pratchett, Small Gods