This is out of the current scope of discussion, but will eventually come to pass; on their own, a "Weblog entry" as they're being modeled here is of limited use; it's just another representation format. What's interesting is when you lump them together in something called a 'feed' so people can poll it, distribute it, etc. and hit that nice sweet spot that PointCast missed and RSS and Weblogs get so well.
I took a stab at initial thoughts a while back. -- MarkNottingham
Maybe it is in scope: the entry has a relation to a channel similar to its relation to a thread. See the discussion at SiteAndSyndication.
[ChristianCrumlish, RefactorOk] I think it is in scope, because the stringing together of entries is important both to weblogs as they are now and to any kind of syndicated-feed model. In fact, I think we should abstract away from the reverse-chronological lockstep and separate out order of presentation from the various datestampings discussed elsewhere. ScriptingNews, for example, does show the most recent *day* first, but the order of entries on a given day are entirely determined by the weblog's author (and managed in, I presume, Manila and Dave's outliner). Likewise, nonweblog publications that may want to use this format - such as newspapers and their analogues - will want to be able to control the order of items without having to game the dates. Finally, even in the world of weblogs, I'd like to see software that can read entry precedence (say, on import from another system) and display entries accordingly.
[GrantCarpenter, RefactorOk] I don't necessarily know that either EntryChannels or SiteAndSyndication is the right place to ask the question (or if the question itself is even in scope), but is there a way to define a core set of attributes that define a collection of WellFormedEntries in such a way that it can be portable across client- and server-side tools? One of the best arguments I've seen for extending RSS is to encapsulate it within meaningful Containers, such that it basically defines a weblog in its entirety. In other words (and I apologize if this veers too much towards being concrete or a specific use case) a base collection of attibutes about a content source in addition to its contained content (WFE's) would make basic bulk portability for tools as straightforward as consuming/producing individual entries.
ChrisHollander Whenver I start thinking about organizing logical groupings of entities, entries, I find myself inexorably tumbling down the rabbit hole or directories, meta directories, and OPML. Basically, I can think of my 'blog as two things: 1.) a completely unorganized bag of every well formed EchoEntries i've ever made, 2.)a heirarchtical organization of smaller collections of well formed EchoEntries. How about each "site" exposes an OPML outline of all of its "channels"? The channels could of course be nested, and of course each EchoEntry could appear in multiple channels. The CMS and the author would be responsible for maintaining that taxonomy/ontology. consumers could subscribe to any level of the organization.
[ZhangYining RefactorOk] The combination of OPML-in-feed and Trackback together might potentially open up a big Trackback Spam:
Trackback enables weblogs the automatic inclusion of contents;
OPML-in-feed enables trackback spammers to easily locate next group of victims (like address books/mailing list for email spammer);
Paranoia can be healthy. This concern was expressed strongly by Baldur Bjarnason. The seed for dealing with this is discussed at Identity.
Building in an outline structure seems too specific for this effort to define Entry. Take a look at Containers and Related to see if there's a foundation for the outline data structure there.