[LibX] cue design alternatives (was: broken embedded cues in Amazon.com)
rossfsinger at gmail.com
Wed May 28 09:02:06 PDT 2008
Ah, right, there are *two* steps to embedding cues, I suppose.
Reading the page for something to cue off of, and then figuring out
where in the DOM to stick something.
This still may be feasible with Dapper or Pipes; after all you're
targeting expected data in most sites: ISBNs, ISSNs, DOIs, etc. You
mostly know what you're looking for and roughly where it would be (at
least, regarding the Amazons, the Google Book Searches, the Google
Scholars, etc.). Serendipitous discovery of cues would be less likely
(although there'd be no reason why it couldn't be discovered the same
way it would be discovered anywhere else), but you'd also be more
sheltered against small changes to the DOM.
I think you can definitely grab the data points you'd want, and
possibly pass something to clue LibX as to what DOM element to update.
Not sure, but it would be an interesting experiment, perhaps.
More information about the Libx