Re: [syndication] Is a Feed the right place for your Data?

[Ken MacLeod <ken@bitsko.slc.ut.us>]

"Danny Ayers" <danny666@virgilio.it> writes:

> It looks to me like there's an ongoing smudging of what a feed
> actually is - having microcontent like dc:description or a short
> review in an RDF data file is a good use of the format; when data
> which is intended to be persistent drops off the bottom of the page
> then it's getting silly. Site summary material is getting mixed up
> with "what my cat had for breakfast".

I'm more and more coming around to the "super-simple feed format"
approach.  A feed is just notification and lead-in to site
index/summary.  Data is linked from the feed in much the same
technique that an article or message is requested seperately from the
newsgroup/mailbox headers in NNTP and IMAP.

In this model, most "syndication modules" should more accurately be
described as "item modules", that data should be in the item on the
web.

Which brings another "Aha!"  When we decouple the feed resource from
the item resources, the item resources become more amenable to being
in different formats.  That's ParticleWave, yes, but the "Aha!" is
that we can get out of the mode of playing syndication format
favorites.  There's no question about who controls X/HTML or RFC-822,
for example.  There's a bonus in there for the RDF folks as well, once
they figure out the "right way" to put RDF in or near X/HTML, it's a
natural extension -- we don't have to come to them, they will be
coming to X/HTML and that's already clearly in their roadmap.

Note that adding data to X/HTML runs into namespace problems.  Problem
already solved: RFC 2731, Encoding Dublin Core Metadata in HTML,
defines a method for using namespaces in X/HTML <meta> and <link>
tags.

 -- Ken