Sam Ruby wrote:
> 
> Simply scheduling all Paces which are not obviously withdrawn, results in:
> 
> = Currently Under Discussion By WG =
>   PaceAnchorSupport

-1. Separation of concerns. If there are microformats people following
who claim they need/want this, I would consider changing to 0.

>   PaceCategoryListing

+0 only because I am unsure of the mechanism proposed, it looks bloated
 (I might come back with questions), but listing categories is valuable.

>   PaceClarifyMediaResourceLinks

+0

>   PaceClarifyTitleHeader

+1

>   PaceDifferentRelValue

+1

>   PaceDontUseContentSrc

+0

>   PaceEditLinkMustToMay

+1

>   PaceExamples05

+0. Since the pace is out of the date, the editors will need to pick
these apart if accepted.

I'd suggest the WG review all examples for implicit or contradictory
specifications. I see James T for one has already done so.

>   PaceLinkIntrospection

-1. Doesn't pertain to APP directly, should be a separate draft.

>   PaceMissingDraftHasNoMeaning

+1

>   PacePreserveForeignMarkup

+1. I think the text will be tweaked if accepted.


>   PaceRemoveMemberTypeMust

+1

>   PaceRemoveMemberTypePostMust

+1

>   PaceRemoveTitleHeader

-0

If we seem to be adding these as we go, ie the number of headers ends up
open-ended I might revert to +1 and refer people to RDF for this kind
out of line metadata.

>   PaceRequestURI

+0

It's a nice feature for authors, but I was close to -1 due to
underspecification for developers. I have some questions:

is the ".atom" postfix a requirement on the server?
is 'href' subject or not-subject to xml:base processing?
are trailing '/' and '#' or preceding '#' significant?
has the spam potential of this feature for SEO been considered?

If this makes it, the editors will probably add impl-oriented notes in
the draft calling out some issues. I speculate that to get any kind of
interop, we'll end up getting the server to inform client show it
constructs URIs (which could reintroduce uri-templates)

>   PaceServiceMayHaveWorkspaces

-1.  Having at least one workspace gives clients and servers something
to agree on. A burden of proof is required here (imo). The argument that
forcing servers to expose a workspace will result in junk interop is not
made. The conceivable cases mentioned are not enumerated.

>   PaceServiceShouldHaveWorkspaces

As above.

>   PaceSparqlLink

0 to the idea. SPARQL is still a W3C WD. For all we know this is the
could take as long as XQuery plus SW community has a history of chopping
and changing query languages (I don't believe this is FUD btw).

-1 to the pace. I have no idea where this feature would go within APP.
Can the author(s) revisit and represent in terms of spec text?

>   PaceTitleHeaderOnlyInMediaCollections

+0

>   PaceTwoPrimaryCollections

+1.

[I wonder if we are going to end up needing 'accept' for collections]

>   PaceWorkflows

-1.

I understand the rationale for removing app:draft, plus I like the idea
of keeping APP cruft out of Atom entries for layering reasons, but in
the absence of some other mechanism (presumably collection based) I
object to taking the feature away due to inelegance.

"app:draft is unnecessary" - I find that absurd given more than two
server implementations. Getting your hands and manipulating entries in
draft state is as close to a universal requirement/behaviour for
syndication as I can think of. I'll refer us back to Obasanjo's point
from last year. We need a way to give the client a list of drafts.
Please propose an alternate mechanism before dropping the only one we have.

"Furthermore, it addresses only a small subset of work-flow problems. "

True. However 1) anyone who wants to worry about workflow problems can
go and check out the WfMC to see the size of that effort, 2) drafting is
far and away the dominant workflow problem on our plate.

>   PaceWorkspaceMayHaveCollections

-1. Rationale as per PaceServiceMayHaveWorkspaces above.

>   PaceWorkspaceShouldHaveCollection

-1. Rationale as per PaceServiceMayHaveWorkspaces above.


cheers
Bill

Reply via email to