Re: Consensus calls (PaceAppModified2)

[David Powell <djpowell@xxxxxxxxxxxx>]

Tuesday, August 29, 2006, 5:44:20 PM, Tim Bray wrote:

> PaceAppModified2: ACCEPTED.  Lots of support, only one objection.

> [...]

> PaceOrderCollectionsByAppModified: REJECTED. One for, one against,
> not good enough.

Whilst requiring app:modified just for the improved sort order
proposed by PaceOrderCollectionsByAppModified might not have been
important enough to gain much support; given the context that we
*have* just required app:modified, it seems silly not to use it for
the sort order.

Or am I missing something? Given that both fields are now mandatory,
what is the justification for using atom:updated rather than
app:modified for this job?

Requiring atom:updated for the sort order prevents sync from ever
working with APP. A compromise might be to remove all references to
sort order, and let extensions be used in the introspection document
to assert which order is being used, but I can't see why anyone would
need an order other than app:modified.

-- 
Dave