David M Johnson wrote:
>
>
> On Apr 24, 2006, at 1:16 PM, James M Snell wrote:
>> Alrighty folks, it's been several weeks. Several implementations are
>> out there being tested. We've talked about it, hashed it over, etc.
>> Media collections are broken. Here's a fresh attempt at fixing them.
>>
>> http://www.intertwingly.net/wiki/pie/PaceMediaEntries
>>
>> The WG seems to be in agreement that the current definition of Media
>> Collections is horribly broken.
>
> Actually, the draft 8 definition works for me. The blog server I'm
> working on, Roller, every blog has a collection of entries and a
> collection of uploaded files. And currently, Roller does not store any
> metadata for uploaded files.
>
I noticed :-( ... ;-)
>
>> The key advantage of this approach is that it works seamlessly with
>> existing *casting style applications. It also resolves all of the
>> issues that have been raised in relation to media collections.
>
> The draft 8 approach supported podcasting too.
>
Does anyone have a comprehensive list of Atom-enabled podcasting
clients? The few I tested that work on Linux did not support using
content/@src for referencing enclosed resources.
>[snip]
> Good. I can still have an entries collection and a media collection if I
> want.
>
> Question: without media-types, how do clients discover which types the
> different collections allow?
>
It may turn out that the member-type element is still useful and
shouldn't be removed as the proposal suggests. However, it would likely
need to be augmented to support the notion of mixed collections.
Another point, consider the case with GData, in which collections are
differentiated based on Kinds of entries (event entries, contact
entries, etc). The "entry" and "media" labels are not sufficient to
describing that kind of differentiation.
In any case, I'm +0 on removing member-type completely. I did so in the
proposal because the idea of a "Media collection" would no longer exist
in the protocol even tho implementors could still have collections that
only contain media entries.
- James
