Blogger automated trackback ping tool
Write a Greasemonkey script which adds (outgoing) Trackback functionality to Blogger. I believe a good solution would be to add one or several links to the publish page, similar to my Blogger ping tool to trig each trackback ping manually.
My present favourite thought on how to accomplish this is by processing the published post's links (including both the optional Link: field and any <a href> tags in the post body), to sum up a list of outbound links. Then, for each link in that list, fetch the pointed to page, processing it for structured announcement of a trackback URI for that page (via the RDF markup standardized for this purpose -- there may be hairs to split about how to behave, in case a page would specify more than one trackback URI -- I'll leave that for further pondering and discussion). Filter out all the URLs from the list which did not yield one trackback URI. For each of the links left, add a "Trackback ping" link on the "post published" page.
Points for style awarded for stylishly formatting this list, relating it to A) the linked URL, B) the <a href>link contents</a>, C) the trackback ping of the resource. I'd appreciate any suggestions and HTML mockups very much; this is a bit of my weak spot.
Additional points of style for an additional list, clearly separated from the first, for the filtered out links (and marked as such) for the links lacking a trackback URI we could find, for human manual visit, to pick up any such trackback URIs (as I'm sure many are only marked up for human visitors to find), and related machinery to add the trackback URIs to these, and have ping functionality for them as well. (The mechanics part of which I will also manage myself without any problems.) Aid on the visiofunctional design, again, much appreciated.
My present favourite thought on how to accomplish this is by processing the published post's links (including both the optional Link: field and any <a href> tags in the post body), to sum up a list of outbound links. Then, for each link in that list, fetch the pointed to page, processing it for structured announcement of a trackback URI for that page (via the RDF markup standardized for this purpose -- there may be hairs to split about how to behave, in case a page would specify more than one trackback URI -- I'll leave that for further pondering and discussion). Filter out all the URLs from the list which did not yield one trackback URI. For each of the links left, add a "Trackback ping" link on the "post published" page.
Points for style awarded for stylishly formatting this list, relating it to A) the linked URL, B) the <a href>link contents</a>, C) the trackback ping of the resource. I'd appreciate any suggestions and HTML mockups very much; this is a bit of my weak spot.
Additional points of style for an additional list, clearly separated from the first, for the filtered out links (and marked as such) for the links lacking a trackback URI we could find, for human manual visit, to pick up any such trackback URIs (as I'm sure many are only marked up for human visitors to find), and related machinery to add the trackback URIs to these, and have ping functionality for them as well. (The mechanics part of which I will also manage myself without any problems.) Aid on the visiofunctional design, again, much appreciated.
posted by Johan Sundström
| 13:07
Many blogs do not include the RDF for trackback. I don't think mine does... all I did was include the javascript instructed by haloscan...
Well, for once I see some actual value in the (admittedly very haughty) attitude "Well, too bad; no soup for you!". It's one of these things I whole-heartedly think should not need any manual tracking to pick up.
All web pages have their own non-standard machine unparsable way of visually expressing "this is a trackback URI". That is of course fine for those who consume trackback URIs manually, out of their own preference. It also might be fine if you only really want trackbacks from those who trouble themselves to scan your page visually for the link, as a form of trackback CAPTCHA.
For trackbacks, unlike, say, the DVD standard (which, to the best of my knowledge, does not standardize how to find track 1...N, so a DVD player could offer to skip all the annoying menu systems we happily have to grok and use today), there is a great standard on how to machine readably semantically mark the relation between a page and its trackback URI, when it has one.
Having those should be encouraged. Lacking them should be, rightly so (indignantly might be a bit harsh), gleefully ignored by this tool, at least one of my own making. I might stretch to make it easier to add a hint to the RDF-less on how to add the needed bits, in case someone contributes effective DWIM code to find unmarked trackback URIs, but remember that a link that just says "trackback", or even "trackback URI" could just as well point to any page on the web, perhaps in a context of explaining the word or whatnot, which does not make that link a trackback URI for the page visited.
Note that all of the above festering does not imply that my own blogs, at the moment, do any of this right. I hope they do, but have not researched it enough to tell with any certainty. Making the tool would kind of test that, too. :-)
I am by no means against including such RDF in my page structure, however everything I know about trackbacks I learned by messing with the haloscan interface. It gave me the auto-install code and that is how my blog works for now... how would I go about including the RDF?
Many parts of this idea are realised in the del.icio.us, pinging, and trackback helper.