[Home] [Groups] - Message: [Prev in Group] [Next in Group]
6511: Re: [MUD-Dev] World Persistence, flat files v/s DB v/s ??
[Full Header] [Plain Text]
From: J C Lawrence <claw@under.engr.sgi.com>
Newsgroups: nu.kanga.list.mud-dev
Date: Wed, 01 Apr 1998 13:49:13 -0800
References: [1]
Organization: Kanga.Nu
On Thu, 26 Mar 1998 01:55:50 PST8PDT
Chris Gray<cg@ami-cg.GraySage.Edmonton.AB.CA> wrote:
> In the area we started out discussing, that of attempting to run
> world-changing threads on the clients, that isn't good enough. You
> need to make the entire transaction that the code does atomic. There
> was an example about this a few months ago, involving two agents
> trying to go through a door. I don't remember the details, but one
> agent could block the second, but that second agent had two steps to
> its progress. The key is that the first step must not be allowed to
> update the DB unless the second step also completes, but the second
> step cannot complete if the first agent blocks the door. Simple
> transactions from the client to the server cannot handle this - the
> entire sequence of steps in running the action must be tentatively
> done, then committed or not, depending on things that may have
> happened in the mean time. I'm not explaining this well...
This is the "Dragon's Dinner" scenario posited by Demos which I just
posted.
--
J C Lawrence Internet: claw@null.net
(Contractor) Internet: coder@ibm.net
---------(*) Internet: claw@under.engr.sgi.com
...Honourary Member of Clan McFud -- Teamer's Avenging Monolith...