@Trond Husø — the datetime attribute can contain dates and/or times. No need to overly complicate things with three attributes when one will do.

@anon — for more information see our article Goodbye time, datetime, and pubdate. Hello data and value.

http://lists.w3.org/Archives/Public/public-html/2011Nov/0098.html

you’re mixing semantics – the time tag only makes sense if the only content within it is a representation of the attribute, ie. it really is a data item albeit with one of far greater relevance than many of the others. I think the big problem with time is, as you suggest, that the attribute must be mandatory with the representation as a backwardly compatible option. Formatting would default to the long format for the browser’s locale. Not sure whether this should be overridable by CSS – what would be the other use cases? or a format attribute. That kind of specification would have made implementation a lot more straightforward and that would have encouraged adoption.


<datetime datetime="2009-11-13T20:00+00:00"> Starting at 8pm on November 13, 2009</datetime>
<datetime date="2009-11-13"> November 13, 2009</datetime>
<datetime time="20:00"> Starting at 8pm</datetime>
<datetime date="2009-11-13" Time="20:00"> Starting at 8pm on November 13, 2009</datetime>


I mean:
In the example you are showing a date inside a time tag.
In the next example you are doing the same, saying that the value in the tag is a datetime.

It would be better if the tag was
and that atleast one attribute was required either date or time
so that you could do:
Starting at 8pm
Starting at 8pm on November 13, 2009
Starting at 8pm on November 13, 2009

I do believe it is a need for a time/date tag in HTML. It is a welcoming suggestion for articles / CRM systems where you can both mark up the time and date of the article, and also time and date of any update.

Keep up the good work with the spec.

Trond

@Bruce – One use-case or a datetime attribute in IMG might be for discoverability: “find me all images from 1 April 2009″; but Charlie’s point about EXIF has merit.