Re: Scenario: Trackbacks

Joseph Reagle wrote:
> [As a preliminary aside, I note that this page:
> http://philringnalda.com/blog/2002/06/a_pox_on_validations_house.php
> which talks about the problem of RDF in XHTML for trackbacks -- and
> includes embedded RDF (html/body/div/RDF) -- is said to be valid?!
> How can that be?

[ [Doubly aside: I'm not ever going to live down that ill-considered,
accidently-published rant, am I?]  Not only is
http://philringnalda.com/blog/2002/08/trackback_and_validation_summary.php a
better intro to TrackBack and validation, since my blood pressure was down,
it's also the genesis of how my site is "valid" with embedded RDF: I cloak.
If the User Agent is one of the validators, I don't include the RDF. It's
cheating, but it feels better than hiding it in comments, where XML parsers
are allowed to discard it before it ever has a chance to see an RDF parser.]

> format: RDF, the present syntax is RDF/XML, though I'm not
> sure this was absolutely necessary.

Not at all necessary in current implementions: at least Movable Type,
probably most implementations, just parse the RDF with a regex, so it's not
RDF so much as fixed-format text (namespace prefixes must be exactly those
in the spec, etc.) which happens to also be RDF.

> o scope: PARTS of XHTML, this is the only scenario I've seen where
> folks are treating "entries on a page" as distinct resources that
> need to be annotated and described in RDF.

Worse yet, it's not actually chunks of XHTML, as it would be if the MT
templates defined <div class="blogbody" id="p<BlogItemID>">, it's generally
metadata about an empty <a name=""></a>, which may or may not be in the
particular page (my main page describes 15 separate per-post pages, someone
with only monthly archives might have a main page describing 15 chunks of
one or two other pages). It's a little less painful to think about if you
think of a weblog's "main page" as being metadata (rather too thorough
metadata) about the archives, so that you have both RDF and XHTML metadata
about the permanent resource, though you still sometimes have to deal with
the problem of describing 15 empty anchor tags in one other page.

> o validation XHTML: it'd be good if the host page is validatable XHTML.

At least from TrackBack's side of the fence, that's not a "good", it's a
MUST. That was the impetus for the mad flapping around looking for a way to
make it validatable last summer: for good or ill, there are enough
influential webloggers who won't do anything if involves invalidating their
XHTML that TrackBack probably would have flopped without the evil
inspiration to comment it out until something better came along.

Phil Ringnalda

Received on Monday, 9 June 2003 19:59:35 UTC