HubLink

HubLink is a prototype of a Blosxom-powered RSS feed and CSS-based semantic weblog that uses TrackBack, autodiscovery and XML web services to form ridiculously easy groups.

How it goes:

If you write comments about an abstract from HubMed and post them to your weblog, whether using a Moveable Type bookmarklet from an individual abstract page or by blogging an item from a HubMed newsfeed, send a TrackBack ping to http://www.pmbrowser.info/hubfeed.cgi

This will retrieve your post, find the link to a HubMed abstract page, decipher the PubMed ID number of the article referenced, then retrieve from PubMed all the data for that article. This data, plus your TrackBack data, will be posted to HubLink. The HubLink page is also an RSS feed, to which anyone can subscribe.

You can set Moveable Type to automatically ping the address above whenever you post to a particular category of your weblog. You can also post to individual categories of HubLink. For example, sending a TrackBack ping to http://www.pmbrowser.info/hubfeed.cgi/malaria will create a group called 'malaria', or add to that group if it already exists. You could subscribe to the RSS feed of that group: http://www.pmbrowser.info/hublink.cgi/malaria

HubLink also attempts to include Dublin Core metadata alongside each article (you can see this by viewing the source). If anyone can clearly describe what this should officially look like, I'll fix it straight away.

I've borrowed many other people's ideas for this and haven't had time to acknowledge them or modify the pages yet. Here are a few links: [1] [2] [3] [4] [5] [6]

Browser compatibility: IE6/Win doesn't use the stylesheet; IE5/Mac shows some hidden elements; Safari v60 doesn't follow clicked links (and you might have to press reload to see everything). Mozilla 1.3 and Chimera 0.6 are fine.

Still to come: