[MUD-Dev] Re: DBMS in MU*'s

[J C Lawrence]

On Mon, 20 Jul 1998 15:50:33 -0400 (EDT) 
s001gmu <s001gmu at nova.wright.edu> wrote:

> (btw, JC, you've won me over on the spoof's idea.  I had decided to
> *cough* borrow the watchers before, and had been deliberating on
> spoffs.  *bow*)

Spoofs BTw have a large weakness: Making sure they are destroyed.  Its
(comparitively) easy to have objects spoofed and to either forget to
destroy the spoof or to leave permutations open that don't destroy the
spoof.  The result is that you get objects accumulating spoofs ad
infinitum (I found one such leak when I noticed that my character
object had over 300 spoofs) with method calls becming slower and
slower as they filter thru more and more layers of spoofs.

Think of it like memory leaks -- it really is the same probl;em in
essence.

To help address the problem I've made frameworks for spoofs, default
constructor, destructor etc, as well as a simple way to state the
conditions that a spoof should exist in, and the conditions it
shouldn't.  This last is a catch-all attempt operating on the basis
that a hidden function gets called whenever a spoof method is invoked.
The hidden function then checks if the current state is valid for the
relevant spoof, and if not destructs the spoof.  Essentially a sort of
rule based garbage collection.  My fear is that user programmer will
come to rely on the rule sets rather than defining their own
detection/destruction semantics -- meaning that errors or absences in
the rule sets provide new leak opportunities.

--
J C Lawrence                               Internet: claw at null.net
(Contractor)                               Internet: coder at ibm.net
---------(*)                     Internet: claw at under.engr.sgi.com
...Honourary Member of Clan McFud -- Teamer's Avenging Monolith...