I'm becoming increasingly suspicious of the MarkDown school of thought that says that we need to use ordinary characters to mark up text with formatting information. I think it came about in the time when all we had was plain-text editing — WardsWiki and co? — so actual richer-text was simply not even an option. But now?
I mean, YES, we need to agree on serialisation of such bits of richer-text, but I'm pretty sure a bit of XML can handle that. What I question is the UI/input side of things. What's the difference between typing " – a word or two – " and typing "ctrl-I – words – ctrl-I"? None. Nothing.
It's also telling that the whole simple-markup movement originates with Programmers. Who work mostly with plain-text editors. Coincidence? I think not.
But to all those Programmers writing tools for people who write words I'll say it out loud (again!): Writers don't want to use MarkDown. They want rich-text editing, and to see the formatting as they intend it on a screen. How you store that in permanent storage, or how you send it across the wires to another computer... that's your problem, and nobody else cares as long as its decodable by other bits of software.
I mean, I do get the "plain-text serialisation" argument, and agree with it completely — our text should be portable between different applications. I don't want my writing work locked up behind some Microsoftian vulture-wall. But a simple bit of XML, or — heavens forfend — a text-formatting subset of HTML, would work just as well for the purpose. If not better.
Bring back WYSIWIG editing. Even our IDEs can handle rich-text nowadays.