asterisks as bold or italic? (another push)

[european bob]

On Tue, 2004-03-30 at 21:45, Timothy Binder wrote:
> > As does OpenOffice.org. But, we're confusing semantic and presentation
> > in the above.
> 
> But I think that existing practices are important. If you have to 
> change gears only when working with Markdown, than it's arguably bad 
> user interface, especially if it's counterintuitive, which is the crux 
> of the argument.
...
> Italics are a more subtle form of emphasis than boldface. This is 
> probably why there was the initial correspondence of <em>:<i> and 
> <strong>:<b>. I would argue that it continues to hold in most cases and 
> is broken only when someone wants to be "edgy" in their design.

I still think you're confusing semantic with presentation. We both agree
strong is stronger emphasis than em. We both agree that ** is stronger
than *. If ** is to be stronger than *, then they need to correspond to
strong and em. If you want _ to be weaker than em, you're basically
stuffed because HTML has no such thing. So you have a couple of options;

If you want _ to be em and * to be strong, then ** is redundant. But it
also means that by default people are mostly going be using *most
emphasis available*, not the _next level of emphasis available_. That's
a bit shouty.

Alternatively, _ could be rendered purely as <i> or <u>. But that's
pretty presentation and has very little semantic.

> But my main point is that underlining has been synonymous with italics 
> since the days of the typewriter. (See any style guide on using titles. 
> Ones that appear in italics in print are underlined on a typewriter -- 
> the original plain text entry system.) To break this link is 
> counterintuitive.

I disagree. I'm afraid you've lost that battle to the hyperlink, as you
suggest. I would also suggest that anyone using Markdown is going to be
far more familiar with a word processor than a fixed-face typewriter
(even modern typewriters could do italic without resorting to
overtyping), and there is no such association within wordprocessors
because we've never had a single fixed face, and therefore italics has
never been an issue.

I think the argument for replacing */** with _/* is a stronger argument
that replacing it with _/*/**. More that two tags doesn't fit HTML well,
and as Markdown is a language for writing to the web I don't really
think it's much of a starter. I'm not hugely convinced of _==emphasis
myself, so I disagree with both the current implementation (_==*) and
your proposal (_==*,*==**) anyway ;) 

_/* I guess wins in terms of Huffman, but */** is more natural in terms
of levels of emphasis to my mind. I also think _/* are closer (in terms
of difference of emphasis) than */**, so having _/* as synonyms seems a
good idea. But then, I don't really know what you're arguing in favour
of with respect to **/__ - would you just dump them?

--bob.