Despite my strong sympathies for SGML, I do see some value in
something very lightweight and richtext-like (although I'd prefer to
see it separated from MIME).
We should probably go for some sort of rich text eventually, but I
always thought that, who was it, Bill Janssen's, I think, idea of a
super lightweight version of richtext that was barely richer than
plain text would be kinda neat, at least in the short term.
What I mean is, using things like *bold*, /italics/ and _underlining_,
and some very simple hard/soft newline handling.
The objection that it is hard to match up pairs of *s is simply bogus.
It is very easy for the implementation to ensure that the emitted byte
stream contains matched-up pairs of *s, and you could even just have a
rule that says that such pairs of *s are always on the same line, even
if you have a very long bold segment that spans lines.
And the hard/soft newline handling bit can be done just like this very
message, where two successive CRLFs indicate a new paragraph. And you
could do hard newlines by indenting like this:
bold
italic
underline
And you could have the byte sequence ":-)" stand for the Unicode char:
263a white smiling face
= have a nice day
And so on. It doesn't have to be difficult. I'd pursue it myself, if
only I had the time...
Cheers,
Erik
PS I'm NOT saying that "richtext" should be changed to be like this.
Rather, this could be a separately registered text subtype,
used by people that WANT to use it, no more no less.