🚀 go-pugleaf

Here's a rather specific question that I have regarding a client-server game
that we're trying to set up (for info on the game, mail 'qgh-request@reed.edu')

I already have a client-server communication system set up that uses
a TCP connection; socket-to-socket.  I have several questions:

(Assuming that the time factor doesn't come in at all :-)
Would it be quicker/easier on the server for me to use a UDP (messaging)
protocol?  Or should I stay with the constant connection that I already
have?

Are there any general tips that anyone can give me on this type
of programming (C) - the way I have it set up in my mind is that
I'd have client code, server code, and an 'independent' library
that both use, which contains all of the communications stuff.
The server acts as a database with security, and the client displays
all the data...

And finally, does anyone have any good code to wait on several
sockets at one time for information?  Basically I want the server
to block on multiple sockets until data is sent...

Any general tips at all are appreciated...  We are going to be
implementing this on a DECStation running Ultrix, and if anyone
has a spare vanilla-BSD 4.3 or a plain SYSV machine I can
possibly use, that would be nice :-)

--Titus
(P.S. - QGH stands for Quest for Galactic Hegemony :-)
-- 
"Never put off until tomorrow, that which can be done the day after tomorrow"
          -- C. Titus Brown, anonymous student, brown@reed.edu

                UNIX is good, you say?  Which UNIX, say I!