[uf-discuss] Rationale for providing hAtom instead of Atom/RSS feeds

[James Aylett]

On Wed, May 28, 2008 at 02:07:29PM +0100, André Luís wrote:

> Maybe rel="next" could be made a "recommendation" to establish links
> between pages containing streams of uf's. Generically.

HTML defines rel='next' [1] in terms of the relationship between the
HTML documents. There's a semantic concern here of the relationship
between the microformat data embedded in the page and the HTML
document. There may be more than one embedded collection of
microformats in a single HTML document (indeed, this is a good thing
:-).

I suspect a generic recommendation here is unwise, as the implication
(paged streams of microformats should be presented as HTML pages
linked together using rel='next') is not reflexive (it is certainly
not true that all HTML pages with rel='next' and containing
microformats that /may/ be presented as paged streams actually embody
a paged stream of microformats; consider for instance a paged preview
of a year using hCalendar to point out conferences; those hCalendar
instances might be in a sidebar and present on every page of the
preview).

The XFN pattern isn't so simplistic, of course: rel='me next' is
different to rel='next'. (Although I'd consider this a slightly
questionable reading of the definition of @rel, since the semantic
relationship between multiple link-types on the same anchor or link
is, I believe, not specified in HTML.)

In the case of hAtom, rel='feed next' would achieve a similar
distinction. A general recommendation could presumably be made that
this approach is the correct one, if it is agreed that this approach
is the correct one.

James

[1] <http://www.w3.org/TR/html401/types.html#type-links>

-- 
/--------------------------------------------------------------------------\
  James Aylett                                                  xapian.org
  james at tartarus.org                               uncertaintydivision.org