Aryeh wrote: >It's kind of a fake, though, since the definition includes "spans of >text whose typical typographic presentation is boldened" and "other >prose whose typical typographic presentation is italicized". With >those semantics, there's no sensible way to render them in any medium >except bold and italics. In speech, you could never present them >properly based on those semantics -- you'd probably just have to >ignore them. For example, even if you wanted to audibly offset >italicized thoughts (which you probably don't), you can't distinguish >thoughts from ship names.
According to the spec the "i element represents a span of text in an alternate voice or mood". It's very easy to do that in speech but very hard in writing. That's why we have emoticons and <irony> tags. The new semantics are pretty solid for i. Admittedly, it's harder to make the case for the b element. b is closely tied to presentation. Its purpose is to "stylistically offset" something. Just like the mark element is used to highlight something in a different context, b is used to highlight something in the original context. In both cases leaving the highlighting out wouldn't change the meaning. b is an accessibility feature which makes it easier to identify key parts regardless of medium. I'd agree that b has the weakest semantics of all the semantic elements in the spec. Using spans with classes would work just as well. Aryeh wrote: >The presentation-independence is hollow: >the semantics are such that it is correct to use <b>/<i> for exactly >those things that are conventionally bolded or italicized. You're implying that these things are conventionally bolded or italicized as an end in itself. Most of the time there's a reason why things are bolded or italicized other than "I don't like regular type". The restricted set of means for conveying semantics in type-setting doesn't mean we can't use a richer set of elements in HTML. Even if at the end of the day all that richness is presented in bold and italics. Google doesn't care ;-)
