Socket Question

[mgarton@xxxxxxxxxxxxxxx]

I have a socket client job that runs in batch with a persistent connection 
to a vendor's RS6000.    The client job just sits and spins until data 
hits a data queue.  My question is regarding error recovery for when the 
connection drops.  Currently, if the client gets an error on the send 
(using Scott's WrLine utility), I end the connection, reconnect and 
resend.  I would like the application to be proactive.  In other words, 
what is the best way to check to make sure the connection is still live 
before I try to send data.  It is often several minutes in between 
transactions and I don't want wait until I try to send data to find out 
the connection has dropped.  I am new to socket programming and this is my 
first "real" application.  Thanks.


Mark Garton