<dialog> element (was RE: Implementor feedback on new elements in HTML5)

On Monday, August 31, 2009 10:56 PM, Jonas Sicking wrote:
> On Tue, Sep 1, 2009 at 2:29 AM, Maciej Stachowiak<mjs@apple.com> wrote:
> > On Aug 31, 2009, at 10:14 PM, Jonas Sicking wrote:
> >>> - <dialog> element
> >>>  This essentially gives the same behavior as <dl> but with
> appropriate
> >>> semantics for logs of conversations. It seems useful and easy to
> >>> implement.
> >>
> >> Useful for what? I don't yet understand what anyone needs this element
> >> for.
> >
> > One example would be markup for a chat log. I'm personally not sure it
> adds
> > a lot of value for this case, but I'm not sure if there is an existing
> good
> > way to mark up logs of conversations, and it seems like a fairly common
> use
> > case.
> 
> What use is marking up a chat log? What value does it add?
> 
> > I don't feel strongly about this, but it doesn't seem like
> > implementing it would be a problem.
> 
> Indeed, implementing <dialog> in a browser is trivial. A feature that
> isn't used, or mostly used wrongly, still adds a cost (spec bloat,
> tutorial bloat, author confusion, name collisions with future features
> etc), my concern is purely that.

This is our concern too but I'd add test cost to the list. Every feature
that we add has a significant test cost making sure it has been added
exactly as the spec requires, testing for accidental regressions, and
adding all the new test cases to our automation tools (plus the time the
additional tests take to run from then on).

On Tuesday, September 01, 2009 3:56 AM, Robin Berjon wrote:
> Yeah, unlike the others this one doesn't strike me as making the cut.
> As currently specified it doesn't see to me to be up to representing all
> but the simplest form of dialogue. You need more structure than that for
> a screenplay, and even a chat log requires things like joins/ parts. This
> seems best left up to a form of microdata.

We also don't think <dialog> adds sufficient value to justify the spec,
implementation, and test cost.

Cheers,

Adrian.

Received on Tuesday, 1 September 2009 14:55:42 UTC