By embedding the RDF into the page, you don’t have problems that people people on something such as blogspot have — inability to control exactly what files are placed on their servers.
]]>I recently just took them out of my gateway because it was reformulated to only include excerpts of a few recent entries. All of the TrackBack info made the page rather fat.
]]>But I am writing more to give the reasons why we went with embedded RDF and not with external RDF files.
In the first betas, we used a <link> tag that pointed directly to the TrackBack ping URL for an entry. This only worked on the individual archive pages, because the ping URL is entry-specific, so it wouldn’t work with multiple entries per page.
So that wasn’t working, because people needed to be able to use TrackBack on non-individual archives. So we instead started using the RDF to represent metadata about each entry. We added this RDF anywhere where people had indicated that they might want to use TrackBack: archives, index, etc. We didn’t really think that adding the RDF could possibly be a *downside*… after all, it’s just adding more metadata about each entry. Metadata that is useful for using the bookmarklet, and could be useful for other applications (maybe).
That’s the reason why we went to using RDF. BTW, we never really thought about using external RDF files, mainly because we already have enough files to write when archives are rebuilt, and we didn’t want to add the headache of managing another set of files. So we made it embedded.
]]>Would the presence of the tb:pingURL then imply that rdf:about is the permalink, I guess?
]]>Yeah, in a single tag like that everything else is describing the rdf:about, so since your descriptions are saying things about the permalink (it has this title, this date, this TrackBack URL):
<rdf:RDF xmlns:rdf=”http://www.w3.org/1999/02/22-rdf-syntax-ns#”
xmlns:dc=”http://purl.org/dc/elements/1.1/
xmlns:tb=”http://example.org/tb/”>
<rdf:Description
rdf:about=”http://philringnalda.com/archives/002329.php”
dc:title=”Just say no to TrackBack in index.html”
dc:subject=”trackback”
dc:description=”Sam also asks if he should have TrackBack RDF in his main page. I say no, for two reasons. First,…”
dc:creator=”Phil”
dc:date=”2002-09-2311:30:35-08:00”
tb:tbURL=”http://philringnalda.com/mt/mt-tb.cgi?tb_id=96” />
</rdf:RDF>
Would parse out as saying:
(/archives/002329.php) (dc:title) ”Just say no to TrackBack in index.html” .
(/archives/002329.php) (dc:subject) ”trackback” .
(/archives/002329.php) (dc:description) ”Sam also asks if he should have TrackBack RDF in his main page. I say no, for two reasons. First,…” .
(/archives/002329.php) (dc:creator) ”Phil” .
(/archives/002329.php) (dc:date) ”2002-09-2311:30:35-08:00” .
(/archives/002329.php) (tb:tbURL) ”http://philringnalda.com/mt/mt-tb.cgi?tb_id=96” .
(picture the words ”has a” in the first space, and ”of” in the second space)
Which isn’t quite perfect, in that the URL ends up as a literal rather than a resouce, but I don’t think you can avoid that while sticking to a single empty tag. I’m pretty sure you would have to say something like:
(*** untested, don’t try this at home, example only ***)
<rdf:Description
rdf:about=”http://philringnalda.com/archives/002329.php”
dc:title=”Just say no to TrackBack in index.html”>
<tb:tbURL rdf:resource=”http://philringnalda.com/mt/mt-tb.cgi?tb_id=96” />
</rdf:Description>
in order to get the TrackBack URL parsed as a resource rather than a literal. Does it matter? Not sure. Ten years down the road, when you want to specify that
(/archives/002329.php) (tb:tbURL) (/mt/mt-tb.cgi?tb_id=96”) .
and
(/mt/mt-tb.cgi?tb_id=96”) (tb:mode) ”Jabber” .
it might make a difference, but I’m not very good at predicting the future that way (more likely, the RDF WG will have changed the syntax completely two or three times by then).
]]>