[MUD-Dev] Re: Dynamic muds
spin at triode.net.au
Fri Sep 24 13:25:19 New Zealand Standard Time 1999
On Thu, Sep 23, 1999 at 03:42:45PM -0700, Ilya, Game Commandos wrote:
> On Fri, 24 Sep 1999 05:28:54 -0700 (PDT), Katrina McClelan wrote:
> ))A typical diku mud merges like items as it is, which is
> ))more CPU intensive, either in pre-sorting or in run time sorting than it
> ))would be to splice text. It's an ugly algorithm to implement in code, but
> ))that has nothing to do with its run time performance.
> I've seen our friend at The Eternal City implement this the
> other way around (and was fascinated by it) -- that is, there
> are group objects. You can look for things in this group. At
> the time someone does that, one particular example of a member
> of that group is created. Until then, the group objects are
> all that exist. And eventually, the individual instantiations
> are subsumed into the group, if possible.
> I have a few ideas to extend this, but that's the core idea.
> Anybody else doing this, or have thoughts about this?
I'm not real sure what the question is here, are you suggesting that all
like items on the entire mud that are not being interacted with are stored
in a group container with some sort of primitive pointer at the location
of the dormant items?
(Hmm, this could be easy enough to try, I guess the
trick is to actually create them from the group container to start off
with rather then trying to match it once created. Whoops, what about
as the objects change, lots of icky matching to see if the entire group
has changed into another group and can be combined. Although I guess that
could be checked infrequently... Argh, now I just have to try it.)
My current mud (diku-based) uses something very similiar to Smaug's
object handling, in that an object may have a count for like items. It
differs in two ways, (1) I didn't standardise my items to increase the
number of like objects, (2) all object functions handle multiples (Smaug
splits each individual item off to handle it). As I have a small player
base, there is very little benefit from this, although creating a
hundred thousand items at once can be amusing.
From the codebases I've seen, before Smaug there was no object grouping,
merely string matching on things like the inventory list, although I'm
sure there were private codebases with such support previously.
I'm currently in the process of grouping mobs, as I think this will make
those terrible polling loops much faster.
\/ I \/
MUD-Dev maillist - MUD-Dev at kanga.nu
More information about the MUD-Dev