2006/2/13, Tim Bray <[EMAIL PROTECTED]>: > I'm in favor of simplification, but there's one scenario that seems > harder to me without media collections. Suppose I'm a cameraphone. > I want to create a post containing one picture and a few words > wrapped around that. How do I do that using APP?
If you can use plain text for your "few words": a. put your metadata in EXIF headers, then POST the picture and let the server extract EXIF metadata b. just POST the image along with a Content-Description header You'll be able to edit the "few words" later by updating the picture with new EXIF headers or an undated Content-Description. If you want HTML or XHTML, you'll have to use what I call a Compound Entry: a. POST a Multipart/Related structure enclosing an Atom Entry and your picture b. POST the picture and let the server generate an Atom Entry for you that you can edit to put your "few words" in (see a previous message from James M about his implementation) Your scenario very looks like the cat-blog scenario, which is the only one that would be easier with a media collection, but only if a media collection is a "binary store" without any metadata associated to its members, which: - is not what's described in protocol drafts since -05 - is not possible if the media collection listing is sent as an Atom Feed I expected my own APP implementation to be finished yesterday evening, it wasn't… I still expect though to have a running implementation 'til the end of this week, with alternate "-04 based" representations (using content negotiation with application/atomcoll+xml media type) -- Thomas Broyer
