[MUD-Dev] Languages for MUD drivers
J C Lawrence
claw at cp.net
Wed Nov 24 12:38:44 New Zealand Daylight Time 1999
On Mon, 22 Nov 1999 14:26:42 -0600
Greg Miller <gmiller at classic-games.com> wrote:
> J C Lawrence wrote:
>> Event distribution assumes that events can be distributed such
>> that minimal synch is required between nodes.
>> Both fail miserably with flash crowds.
> Is this necessarily true? I've been trying to find ways to extract
> enough information to perform scheduling in such a way that events
> operating on the overlapping working sets don't run simultaneously
> any more than is necessary. Of course, that frequently requires a
> way for coders to specify affected objects manually to allow the
> scheduler to figure out what events are compatible, and some may
> not bother... but is that actually worse than other
> (non-event-based) methods of handling distribution?
Its possible. My first reaction is that something along the lines
of Lambert's ahead looking lock detection methods might be able to
help here (I don't recall them in enough detail to comment well, tho
they're in the archives (gotta get some work done)).
The problem of course is:
500 players gather at the castle for a public demonstration (as
happened on UOL).
They can all see each other.
They can all interact with each other.
A singificant percentage of them set off fireworks or various
forms of highly visible spells/effects at roughly the same time
(ie very high event rate, very large working set for each event,
high interaction rate between events).
J C Lawrence Internet: claw at kanga.nu
----------(*) Internet: coder at kanga.nu
...Honorary Member of Clan McFud -- Teamer's Avenging Monolith...
MUD-Dev maillist - MUD-Dev at kanga.nu
More information about the MUD-Dev