> > +:userID # user arrived
> > -:userID # user left
> > s:userID:text # user said something
> > But then, is JS any good at text parsing like this?
>
> Sure.
>
> m = somestring.match(/([+\-s]):(\w+)(:(.*))?$/);
> action = m[1];
> userID = m[2];
> text = m[4];
Though Javascript could do this, as you say, no problem, it would mean the
file loaded wasn't javascript.
It'd be text/plain, and then javascript would analyse it onload.
Doing this once a second wouldn't work out as an advantage, I don't think.
I also don't see much improvement between:
+:userID
-:userID
s:userID:text
and:
addUser('userid');
rmUser('userid');
addMsg('userid','string');
I know it was me that said you want to keep the bytes per refresh down, but
I don't see much point in doing it if it adds even more processing at the
client end each second.
G
