UPDATE 9-Nov-04 @ 21:43 GMT: originally forgot to include the clickable nodes that Compendium generates… so now the nodes in the image below are properly clickable, and take you to the original blog entries.

The painfulness of both the manual and diagrammatic representations of the ‘duelling blog thread’ I referred to in the previous post prompted to have another go at doing it more cleanly. I downloaded Compendium (which requires a prior MySQL download to support the database) and quickly whipped up the following, all of which include dragged-and-dropped nodes directly from the websites/blogs/comments in question, and all of which are clickable to launch those aforementioned sites… cool, huh?

Compendium also generates an XML file complete with details of each node, link, position, and contents. A typical node looks like this:

Cute, huh? And it only took me a few minutes to create that entire diagram. I had to label the arrows, but that took a few right-mouse-clicks to choose from among a pre-stored set of links.

As I just posted to Bertrand Sereno and Simon Buckingham Shum (who heads the Compendium research here in KMi), “it was a little tedious having to relabel each node through the properties dialogue… the default is for the node name to come up with a long-ugly URL… whereas it would be nice for some shorthand to be gleaned, say, from the page title field… BUT OVERALL it was *WAY* simpler to do it this way than by any other method…

*SO*, the interesting thing for a user would be how to semi-automate this process… i.e. have all that diagram ’seeded’ somehow via a start node and ‘trackback crawler’… which obviously couldn’t label the links… *LABELLING WOULD BE THE USER’S JOB*… that’s precisely the point!!!! If everything else could be ’sucked in’ automatically that would be awesome…”

Onward and upward…

Responses are currently closed, but you can trackback from your own site.