IRCCloud solves everything you just mentioned.

The key is in the clients, not the protocol.

> IRC has deficiencies that ensure that only a small cabal of power users will endure it. I prefer to be part of communities that are more accessible to more walks of life than the person who was bothered to install irssi on his spare EC2 instance.

I've been in, and now run, a channel for about twenty years that, at its peak, had about a hundred active users.

Most of those people were normal folks, from all walks of life, many of whom barely knew how to use a computer.

I would argue that it's only power users who want chat history in the first place. Normal people don't want to go back and read god-knows-how-many lines since they last logged in. They just want to chat.

Offline messaging is a valuable feature for everyone and can be improved by services, without needing a bouncer. Many networks have a "memoserv" that does this. Improve the UX of something like that (should be transparent, no different from a normal pm) and you have good offline messaging.

"I prefer to be part of communities that are more accessible to more walks of life than the person who was bothered to install irssi on his spare EC2 instance."

Then why argue for IRC to be updated with features? Why not go somewhere else and leave the cabal alone?

I'm of the pinnacle of chat crowd, and I will fully acknowledge the difficulty to someone that is not on the technical side. However, popularity does not determine quality. From a programmer (bot/client) creator, its perfect. The protocol is simple, light, and fast. For basic usage there is no confusing things, and even where I've had confusion (DCC), it turned out to be simple. From the view of just a user, I still prefer IRC. The clients are lightweight, my Hexchat client is taking up 28mb of ram right now, and its one of bulkier ones. The lack of images allows compactness and simplicity.