[MUD-Dev] MMORPG/MMOG P2P design

Crosbie Fitch crosbie at cyberspaceengineers.org
Thu Feb 27 12:57:48 New Zealand Daylight Time 2003

>From Felix A. Croes

> Assuming that a human-equivalent solution of the "trust" problem
> exists, it does not follow that the resulting P2P system is suited
> for running a MUD.  Among humans, the sky doesn't fall when
> someone tells a lie.

If you allocate 'ownership' according to both 'interest' and
'trust', then the objects that are most 'interesting' to most
players are owned only by the most trustworthy.

> Also, humans sometimes resort to consulting a central authority to
> resolve conflicts, something a P2P system cannot do by definition
> (I am not just arguing semantics here; introducing a central
> authority to a P2P system may introduce a fatal weakness).

Even humans have realised that a central authority is not
scalable. We use hierarchies. We have local authorities, regional
authorities, state, national, continental, international, global,

I'm not really championing an extreme/anarchic P2P system, but a
distributed one. The problem I'm trying to solve is scalability. You
can still have 'control', but it can be distributed hierarchically,
and not on a crude level of saying "This is the central server,
these are the secondaries, and these are the ISP co-located ones,
etc.". Let the system figure it out, it's not hard for a million
computers to figure out which of them should be at what level in the

> You've answered my three questions on a case-by-case basis, rather
> than with a cogent trust model.  Yes, I noticed that too. :)

> Creating such a model, while maybe not equivalent to creating AI,
> is definitely a hard problem, and one you have to solve before you
> can get anywhere with P2P MUDs.

I reckon it's soluble, but I appreciate few would believe it until
they saw it. That's why I try to show by induction that if humans
can cope, then so can computers.

MUD-Dev mailing list
MUD-Dev at kanga.nu

More information about the MUD-Dev mailing list