no, Stomme poes, you are not the one i was pointing at, no way! i'm sorry and i apologize if you felt that way! you never posted something that made me feel uncomfortable!
You confuse me sometimes, but you don't make me uncomfortable. But I wondered if I was making YOU uncomfortable. : )
as for anchors and why i brought them in our argument. you said that dl provides a way to define a relationship between two elements. i've said that in html the anchor mechanism is the one doing that. then why not use anchors for defining the relation between "Q"s and "A"s? because it's just not the way to do it.
Ah ok, except hyperlinks (usually) link to documents. While I have used links in-page to "link" related things, it was special-cases: A note with a *, which in print means "read the footnote"... and I've made those links to thepage.html#footer so readers of long pages can skip to that footnote easier (mostly in terms&conditions pages).
But if the answers weren't on a separate page, I wouldn't use anchors. More than any actual implied linking between words, they are more known for causing browser behaviour.
html semantic disregards text purpose or meaning. you imply that html should look for inside text relationships and outline them. not possible.
I believe HTML should, when possible, reflect the text relationships. If you were putting dictionary definitions on a web page, you would use a dl... not because it's especially easy to manipulate into a form that makes sense (visually) on a web page, but because that tag was created in order to reflect the same key-value pair relationship that terms and definitions have.
I believe HTML can only do this very loosely and in general, and that it can't go any further into semantics without either (real) XHTML or microformats, one of the two. But, HTML markup does indeed express relationships. You're right that it can't tell WHAT the text is, but relationships are implied... same as a header above some non-header content implies that the non-header content has something to do with the text in the header (even if HTML doesn't know or care what it is).
also, there are a series of issues that cannot be resolved with the use of dl regarding the subject of Q&As: when Q and A are more than inline, and neither fit in a dt, when Q is on one page, A is on another and probably other.
Yes, there can be various circumstances which would tell the page author not to use a particular method. In your first example, a dl would not do what the author needs; in the second, the anchors you mentioned would actually be perfect.
and strictly from dl point of view, an answer is not a definition for a question.
Deathshadow takes this view too: he will only use a dl for pure terms and definitions, instead of stretching the meaning to "key-value pairs". Which is fine... as an author, you determine what markup semantics you think you can use.
The OP asked what the best markup would be. Most of us said one of two things would generally work: a dl, or header-content pairs. Obviously if the OP's also doing some other weird stuff like multi-part questions or separate pages then neither of those answers would work all that well.