home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: alt.irc
- Path: sparky!uunet!spool.mu.edu!agate!boulder!ucsu!spot.Colorado.EDU!frechett
- From: frechett@spot.Colorado.EDU (-=Runaway Daemon=-)
- Subject: Re: yes or no? (bot deopping)
- Message-ID: <1992Dec16.105420.5480@ucsu.Colorado.EDU>
- Sender: news@ucsu.Colorado.EDU (USENET News System)
- Nntp-Posting-Host: spot.colorado.edu
- Organization: University of Colorado, Boulder
- References: <85220@ut-emx.uucp> <1992Dec14.022512.20210@fwi.uva.nl> <1ghdrlINNsu6@tik.vtt.fi>
- Date: Wed, 16 Dec 1992 10:54:20 GMT
- Lines: 18
-
- I think I and sum up the problem here..
-
- netsplits are evil. netsplits will always lay waste to the best thought out
- schemes. With centralized control of ANYTHING consistency is maintained....
- for anything that is currently talking to the central authority. But you lose
- overall workability and for those that can't contact the central server
- directly.. they are just screwed. If you distribute your information then
- it can become fragmented and collisions occur.
-
- I just wrote an irc server from ircII script and many of these thoughts
- have crossed my mind in the last few days. Life is so much simpler with
- a single database.
-
- I think I'll start working on redundant networks on a smaller, more easily
- testable scale shortly. Applying some of the knowledge gained from the
- ircIId server, I'll try to implement a mesh DCC channel. We'll see.
-
- ian
-