Actually the discussion was about what models of audio conferencing to use. Noone claimed to have the answer for what solution to use, since that subject is hardly even touched. This is still about what model for initiating conversations, not what solution to use.I know :D
For the people that didn't want to read those 20 (or so) mails;
Tijl proposes a solution with a client/server model where you have dedicated servers, clients that can also acts as a server (the majority of entities on the network) or lightweight clients that are only a client (for example PDAs). Main advantage is that you don't need to have access to a dedicated server, your own PC can host a conference if you have the CPU power and the bandwidth. Problem with dedicated servers is that those probably won't be publicly available because of the required bandwidth/processing power.
Richard is advocating a peer to peer model where all clients establish a direct connection with all other clients in the conference. This way you don't need a server which eliminates potential 'single point of failure' problems.
Also if you've followed the discussion, you can see neither me nor Richard is trying to convince the other is wrong (quite to the opposite in fact). Rather we are are both trying to explain the strong and weak points of the different models that could be used.:D You're both taking your time in doing that ;) IMO the two of you just have different needs/requirements. In those situations it's no use explaining strong/weak points. IMO those points are fairly obvious. It's just what we value more.. Do we want a low(er) bandwidth solution (client/server) or do we want something that can cope with dropped participants in a simple way (p2p), etc. IMO the two of you will never agree on that ;)
Bart
_______________________________________________ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
