Re: PaceDateKenMacLeod (was Re: PaceDateSamRuby +1)

[Antone Roundy <antone@xxxxxxxxxxxxxx>]

On Wednesday, August 11, 2004, at 03:54  PM, Ken MacLeod wrote:
> Neither of PaceDateAsbjornUlsberg or PaceDateAntoneRoundy specify
> interoperable behavior* -- why a producer would include such dates or
> what consumers will do when they receive them.  Both of
> PaceDateSamRuby and PaceDateKenMacLeod define their dates in the
> context of what they do in Atom tools, not just what they "mean".
Why should the behavior be specified or suggested?  Isn't that entirely 
up to the implementor or the user?  I don't think we really need to 
speak on the issue of sort order.  Sometimes people will sort by date, 
sometimes by another criteria.  When sorting by date, if we have more 
than one Date Construct, which to prefer for sorting is an individual 
preference.

As for suggesting that entries MAY not be displayed until a post-dated 
date is reached, I think that would be an appropriate use for 
dcterms:available.  Unless we explicitly bake that semantic into a core 
element, it's anybody's guess whether post-dated entries are intended 
by each publisher to be handled that way.

The single atom:d seems to carry some weak ("MAY") SUGGESTIONS for 
behavior, but doesn't appear to SPECIFY behavior.  And the MEANING of 
the date is not very specific.

The only behavior I can think of that might need to be suggested is 
that the "subjective" date might be thought of as a "display" date.  
But even that would be nothing more than a weak suggestion--some people 
may prefer to display one of the objective dates.