[Standards] Lessons not learned from IRC

Carlo v. Loesch CvL at mail.symlynX.com
Mon Mar 19 13:42:39 UTC 2007


Daniel Noll typeth:
| IRC failed precisely *because* it doesn't globally federate.  They never tried 
| to (mainly due to politics; admins on one network didn't like admins on 
| another network, which is somewhat important for harmony of an IRC network.)

your starting point is somewhere a decade after the invention of IRC,
thus missing the background.

IRC was originally intended to federate, but out of political and
issues it couldn't. See the story of the end of IRC federation at
http://about.psyc.eu/IRC#The_Battle_Of_Eris_-_The_War_Of_IRC

But that was not the point I was talking about. Cite my statement
completely and you'll see..

> you can't forever throw bandwidth at the redundant messages and
> inefficient distributions. irc already went that same path, and
> look were it got.. a thousand scattered subnets.

IRC also failed to scale as an oligarchic network as more and
more users joined the network, the overhead of the global user database
went over the top of what the network links were capable of carrying,
let alone the totally useless costs. Resulting in the notorious
netsplits. In at least one case a netsplit was even institutionalised:
since Europe and North America couldn't coexist in the same network
anymore, IRCnet split off EFnet.

All of these shortcomings of IRC were already known in the early 90s.




More information about the Standards mailing list