In fact, it applies quite well; however, as I've mentioned in the past, I
have some reservations about DIS itself. I think we should be inspired
by it, not adopt it.
> Last time I looked there were 26 different types of PDUs.
... which is one of my objections ...
> It can also contain
> information on articulated or attached parts (i.e. a turret on a tank or a
> blade on a windmill).
... and there's another one. Tesating any moving part as the conceptual
equivalent of a tank turret may not be the best design.
> In a DIS simulation the Windmill would be an entity controlled by one
> process. The process would transmit Entity State PDU messages to describe
> the state of the windmill.
Yes, exactly.
> All the participants could then compute the position of the
> windmill's blades at any given time.
Right, using the dsad reckoning approach. My paper suggests generalizing
this to deterministic functions of time.
> In the Fred and Barney example we have tmese entities: Fred, Barney, and the
> wallet. Each of the entities have a controlling process that transmits
> Entity State PDUs. When both Fred and Barney reach for the wallet the
> process controlling the wallet decides which way the wallet goes. This is
> esslected in the wallets Entity State PDUs. So from the position of the
> wallet Barney knows that Fred picked it up first.
Yes, that works. In fact, I have yet to hear any alternative approaches
that (a) scale, (b) don't require some form of "undo", and (c) don't sat
bandwidth.
-- Bernie Roehl University of Waterloo Dept of Electrical and Computer Engineering Mail:[email protected] Voice: (519) 888-4567 x 2607 [work] URL:http://sunee.uwaterloo.ca/~broehl