Re: A few thoughts on client/server in multi-player games

[Tero Kivinen <kivinen@joker.cs.hut.fi>]

Petri Heinil{ writes:
<lots of stuff about counting steps removed>
> so 3 steps. Ie. the parsing ascii is about 23/3 = 8 times
> slower than using net-host conversion direct. 
> And there should be able to transmit 121 (11*11) points
> at 10 fps. Every point have 2 numbers so the comparision
> is 2783 vs. 363 steps. This might be singnificant.

In current workstations that would be about 55 us vs 7 us, so with
ascii we can only 1796 users with 10 fps... so that won't be any
issue. 

> And note from space. The size of trafferred data is
> overall unsignificant, but it makes the protocol somewhat
> faster in the packets fits into one IP packet. There
> is no need splitting and rerouting the another packets.

Usually the network overhead etc will be much higher than difference
between ascii and binary, so I think we should use the one that is
easier to implement.
-- 
Tero.Kivinen@hut.FI              	     Work : +358-0-451 4032
Magnus Enckellin kuja 9 K 19, 02610, Espoo   Home : +358-0-502 1573