[DGD] Is self nesting ALWAYS a bug?

[Noah Gibbs]

  There are MUDs where self-containment is allowed, and that fact is a feature.  JC Lawrence had a test MUD with some neat semantics for self-containment and seeing "infinite" (many levels, visible to the end user) objects.  In addition to self-containment, the same MUD allowed mirror objects of various kinds which showed inside other objects, and had lines of sight.  So in addition to self-containment, you could also have a room with mirrors on both sides, showing a large number of you and your room's contents.

  I don't remember whether he had a fixed cutoff for the number of repetitions of this that he showed, or if he just detected cycles and did the textual equivalent of "... and so on".

  So whether it's a bug depends on your MUD.  Doing an "EscherMUD" thing isn't for everybody.  But it isn't guaranteed to be a bad thing.

  However, that doesn't mean you need to allow it.

--- On Fri, 3/20/09, Shentino <shentino at gmail.com> wrote:

> From: Shentino <shentino at gmail.com>
> Subject: [DGD] Is self nesting ALWAYS a bug?
> To: "All about Dworkin's Game Driver" <dgd at dworkin.nl>
> Date: Friday, March 20, 2009, 5:32 PM
> Me and someone else who I am choosing to keep anonymous at
> this time just
> got into a huge debate of whether cyclic containment is
> ever a legitimate
> circumstance.
> 
> Personally I think that allowing an object to contain
> itself, even
> indirectly (or similiarly, be a detail of its container, or
> be contained by
> its own detail) will cause no end of anomalies, infinite
> loops, logical
> absurdities, and whatnot.
> 
> But based on runtime overhead, there are concerns over
> whether a trace-up is
> a good investment of ticks just to prevent this.
> 
> Opinions?
> 
> Kotaka btw has self containment outlawed at the core, and
> not even an
> administrator is allowed to override it.  This is in large
> part due to the
> fact that kotaka's world saves are based upon cascading
> the save downwards.
> Phantasmal, however, is flat and based on object numbers.
> 
> I'm thinking though that cache-hot roots would minimize
> a good part of the
> overhead.
> ___________________________________________
> https://mail.dworkin.nl/mailman/listinfo/dgd