Trying to explain why Emacs wiki worked in the end. I think it is an example of how wiki can serve as a persistent memory for an IRC channel.
I consider the EmacsWiki a success at the moment. Why did it succeed?
A wiki is better than a list of channel logs, because information is extracted and polished when it is published on the wiki. The posting of raw logs is practically never done, except for jokes and quotes. This again allows the #emacs community to manifest itself on the wiki. It created a wiki page of the #emacs members , their locations on the globe , and even demographic info .
More about fsbot:
Fsbot doesn’t just expand abbreviations (Interwiki links, like the wiki bot on #wiki). It actually knows the pages:
<kensanata> , iswitch <fsbot> try: IswitchBuffers UseIswitchBuffer <kensanata> , iswitchbuffers <fsbot> From memory, IswitchBuffers is  at http://www.emacswiki.org/cgi-bin/wiki.pl?IswitchBuffers <fsbot>  a package to do quick buffer switching. Use M-x iswitchb-buffer to try it, then add (iswitchb-default-keybindings) to your init file.
As you can see, it will list all entries matching a pattern, if there are more than one, or print the result if there is an exact match. Every item can have several entries; if the item is a page on the wiki, one entry will hold the URL. These items are updated automatically on a regular basis.
See also: WikiIrcGroupForming, from 2005-11-28.
To solve the problem, I think: Maybe some sort of rolling message exchange at the top of any page, optionally bridged to an IRC channel via bot.
Messages could fade out from the all-pages log after, say, 10 minutes.
That is, at the top of every page, you can see the last 5 messages that were chatted. You could click to get a larger log, you can enter a line of text to add a comment. Only major concerns are anti-spamming, which I would leave to later.
Perhaps some sort of integration with the DatePage.