[Reality] Autonomous Events

Tenth tenth@twistedmatrix.com
Thu, 23 Dec 1999 09:19:48 -0500


> > Every
> > 30 seconds is QUITE enough, and you get the impression it's really
> > quacking a lot that way, without cluttering up the event window and
> > spewing 10x as much bandwidth.  It's always been sufficient (in the
> > games
> > we've worked on so far) to do something else.
>
> well, you can do it in LambdaMOO....;-)

	*sighs and shakes his head* Yes, you sure can, and it is documented
quite well in the many "WHY IS LAMDAMOO SO M%(*&@^@F(*%&@ING SLOW?" FAQ lists
devoted to the subject.  ;-)  Allowing every dumbass with a build bit to make
all his code run every three milliseconds is, by far, Not the best way to get
your M** to run smoothly.

	~30 seconds is a nice increment, because even with fairly nasty
internet lag, all the players involved in (whatever your event happens to be)
will have time to read it and react. It also helps to ensure that even if your
authors get hyperactive with timed events, it will take them quite some time to
bog down your system, and even then, the individual event-driven things won't
be quite so annoying.  ;-)  If you really think ~30 is too long, you can always
handle the event twice... I did this by accident the first time I made a
delayed event handler (The RoachMove.java/BrassWind.java set in the /demo
section of the current distribution) and it caused the event to go off about
~15 seconds, although the distribution of the handlings was sort of random...
But then it was a cockroach simulator, so it seemed to fit.

	(And you can always get a ~3 second cycle by setting it to 0, but if
Lamda taught me anything, it was that most things that happen that often tend to
be annoying.)

- Tenth