client-server

[linas@linas.org]

It's been rumoured that Derek Atkins said:
> Lo and behold, there were two TCP connections (one for the
> HTML, and one for the GIF).  They were not concatinated together,
> even though supposedly both the client and server support it.

actually, this is something the netscape navigator does, I don't know
why.   If you looked at a web page with 5 gifs, you'd still have two
connections: the html on the first, and all the gifs on the second.
I think the theory is that it starts requesting gifs *before* its
finished receiving the whole web page, ergo two sockets.

If you use lsof, you *might* see the socket hanging open.  I think
most high traffic web sites shut down open sockets within seconds, so
that they don't pile up huge numbers of open tcp ports.   Also,
I think navigator closes its connection after the last gif, because
it has no reason to keep it open (viz, the next hyperlink click 
is probably to a differnt site anywasy...)

> Sure, HTTP 1.1
> supposedly allows for multiple request/responsed, but in practice I
> have no idea if anyone supports it.

yeah, it works, netscapoe enterprise server really does it, and i'm
98% sure that apache does too.

> "message passing protocol".  And what we're trying to do is more than
> just passing around messages.  

Is it really?

(ok, yes, the event-notification part does require a bit more than
http can simply provide).