PaceAppCategories [was: Consensus calls]

[James M Snell <jasnell@xxxxxxxxx>]

The bug can be addressed by eliminating the fixed and scheme attributes
on out-of-line app:categories elements.

{{{
appInlineCategories =
    element app:categories {
      attribute fixed { 'yes' | 'no' }?,
      attribute scheme { atomUri }?,
      (atomCategory*)
    }

  appOutOfLineCategories =
    element app:categories {
      attribute href { atomURI },
      (empty)
    }


The app:categories element MAY contain a "fixed" attribute, with a value
of either "yes" or "no", indicating whether or not the listing of
categories is considered to be a fixed, or closed set. Collections that
indicate a fixed set MAY reject members that include categories not
specified in the provided listing. Collections that indicate an open set
SHOULD NOT reject otherwise acceptable members that include categories
that are not specified in the provided listing.

The app:categories element MAY contain a "href" attribute, whose value
MUST be an IRI reference identifying a Categories Document. If the
"href" attribute is provided, app:categories MUST be empty and MUST NOT
have the "fixed" or "scheme" attributes.
}}}

- James

Tim Bray wrote:
> [snip]
> PaceAppCategories:  This is the hardest to call.  There are a ton of
> positive and neutral comments, but three -1's. We admit to prejudice in
> that the +1's and +0's include a lot of our implementors.  We think that
> one of the three -1's, from Andreas Sewe, could be turned around if the
> bug he reported in
> http://www.imc.org/atom-protocol/mail-archive/msg05780.html could either
> be fixed, or Andreas convinced that it's not really a bug.  Finally, we
> didn't see any of the opponents arguing that this would be damaging, but
> this Paces's supporters are pretty excited about it, think its lack will
> be a big hole in APP.  So...
> 
>   ACCEPTED.  The WG is asked to comment on Andreas' bug report and
> either refute it or suggest fixes.
> [snip]