@surreality
Damn, I was testing it. Some clients use /<stuff> in other manners like BeipMU. I tested it fine in MUSHclient though...
I have to wonder if doing something unprecedented (for standard MU*s, anyway), like, instead of handing out the address to use standard clients, force people to connect through a web-based client (I can't recall ever seeing anyone try to do it this way)? PennMUSH offers Banana (assuming you can get it set up), there are others like ChroMUD as well. That way, whatever funky code you use for your game to make it accessible with IRC-mirror commands, and such like that, is always consistent to THAT GAME SPECIFICALLY?
It doesn't get all the things on @faraday 's list, but it gets at least a 'consistency' down so that you can start trying to work around some of the differences in commands via softcode command mirroring.
The other thing to consider, MU* works for what we do (those oldbides of us) because it's NOT tied to something like a GUI. Which then would require even MORE code knowledge to make something tie to buttons, make the buttons show what they do, etc. Right now you can go 'Oh, the syntax is snarfblatt'. But it'd be more difficult to create a GUI, unless the client is written to map the buttons in an easy way.
@Sparks
Re: github/coding live. Eh...? I can't recall the last time that games I played on, or games I was running, did 'coding live'. If a game has a dedicated coder, and that coder has a lick of sense, they're going to have a secondary port with a copy of the game to test their shit on without potentially borking shit.