|
BURNINGBIRD
|
|
a node at the edge
|
|
July 02, 2002
 discussion thread
Working on ThreadNeedle's vocabulary tonight. One additional level of sophistication could be to record a posts entire parentage within the RDF e.g. a - b - c The "path" to 'g' would be: a - x - f - g This isn't complete discussion, but is complete thread. Nested data such as this isn't trivial within RDF, but doable. First case of a simple RDF file at http://weblog.burningbird.net/threadreplies.rdf. By using bagID, I should be able to encapsulate each level into a single reified statement that allows each nested level of blogging reply to be processed individually; the bagID prevents recursive looping back on the property. However, the complexity is increased. True, the generating and parsing of the RDF is automated, but I don't want to add unnecessary CPU cycles to the apps. RDF people in audience - comments? Am I cracked on this one? As an aside, line breaks generated by blogging tools are a pain in the butt. HTML break annotation is added to the RDF, which breaks the RDF processors. The only way to avoid this is to have users add their own line breaks (and won't that go over big); or to generate the RDF to be copied and pasted as breakless content - friggen long lines. Most likely break something, and even if it doesn't - solution is inelegant and offensive. A better bet would be to have a tool pre-process the RDF and pull out extraneous HTML garbage. Same tool can also grab multiple RDF blocks within same document - something many of the RDF tools don't like. Unfortunately, this puts burden on those building tools to process ThreadNeedle data directly from files. No solution yet to the problem of how to distribute RDF so that no dependency exists on ThreadNeedle. Or I should say, no solution yet as to how to track the distributed bits of the discussion within several different weblog postings without reliance on ThreadNeedle. This isn't necessary to first release of ThreadNeedle - but bothers me nonetheless. (I desperately need DSL - this work over a modem is slow torture.)
Posted by Bb at July 02, 2002 02:39 AM Comments
as per Joe Gregorio's intro http://bitworking.org/2002/07/01.html), and David Weinberger's write-up, Very pleased to meet you! Jon, thanks for stopping by! I'm in process of developing initial prototype at this moment. Posted by: Bb aka Shelley aka Weblog Bosswoman on July 2, 2002 11:56 AMBb, if you build a good thing and spec it well enough tool makers will do the work to integrate. It makes sense from a developer perspective to give our users as many options as we can. I wouldnt expect blogger support, of course, because as I understand it Ev has his hands pretty full ;) Posted by: ruzz on July 2, 2002 02:51 PMRuzz, I'm hoping that eventually the tools will generate the RDF for embedding. Ev _always_ has his hands full, poor guy. Posted by: Bb aka Shelley aka Weblog Bosswoman on July 2, 2002 05:55 PMWelcome to my modem hell. I spent yesterday at the computer lab at school ... dammit if I wasn't weeping when I came home to my pitiful dial-up... Posted by: shar aka sharon o aka me on July 3, 2002 11:38 AMUnfortunately, with the ThreadNeedle effort in particular, modem just don't hack it - especially Earthlink which oversells its access. Posted by: Bb aka Shelley aka Weblog Bosswoman on July 3, 2002 09:23 PMBb, looking at your sample RDF, I'm not sure what you're doing in a few places. You use "thrdndl:thread" and "thrdndl:replyTo" in rdf:resource and rdf:about attributes that take URIs, which implies the existance of a thrdndl: URI scheme. I'm pretty sure that's not what you mean. Posted by: Dave Menendez on July 4, 2002 01:22 AMDave, you're right, I should be using the RDF-URI encoding. Got lazy. Posted by: Bb aka Shelley aka Weblog Bosswoman on July 4, 2002 04:14 PMPost a comment
|