[ct-user] serial nr sequencing - an other view
Ray Goff" <firstname.lastname@example.org
Sun, 3 Jun 2001 15:13:20 +0100
I whole heartedly agree with Zoli, I have also written a CT compatible
logging programme which runs in a similar manner simply because we were
being penalised for sending incorrect serial numbers in one contest we
enter. I would love to go back to CT for that contest, but until the serial
number problem is licked its not worth the risk. We found that we were
either sending duplicate serial numbers or higher serial numbers for earlier
I my design the PC's are linked together on an Ethernet network, with one PC
is the Server and the remaining PC's running as clients. Each client
requests serial numbers from the server and reports back contacts to the
server. The server updates all the clients with all contacts made and the
latest serial number in use.
Each PC updates its local copy of the log to reflect the contact made by
other PC's on the network. We even leave an extra PC on the network logging
all the contacts firstly to be sure we didn't miss any and secondly to let
visitors look at our score.
I accept that the server becomes the single point of failure in the system,
but at least the serial numbers are correct. If the server does go down,
then each PC immediately notices the broken network connection and runs
standalone until the server is brought back on line.
Unlike Zoli, our server PC is not designated as the run PC, it is simply one
of the PC's on the network, usually the one with the best specification.
Administrative requests: ct-user-REQUEST@contesting.com