Progressive Feed Filtering through Hierarchical Tag Grazing

The major advantage the Social Bookmarking system Raw Sugar has over del.icio.us right now is the hierarchical organisation of tags. To capitalize on that I'd like to see Raw Sugar supplying those hierarchies in OPML format with the accumulative Tag feeds at each node level.

To explain what I mean I've manually built this hierarchy -

Internet -> Software -> Programming -> Ruby -> Ruby on Rails

And you can graze it at Grazr.com or Optimal Browser.com. At each level we see an indication of the number of feed items corresponding to the tag. For instance, at the top level Internet has 56,984 pages of results, Software has 10,082, Programming has 5,199 and so on.

What we can see is that as we drill down through the hierarchy we are progressively filtering the volume of items as we zero in on our specific area of interest. Optimal is probably the better of the two grazers for this purpose right now but what we really need is live feed updates a la LiveMarks. Imagine seeing new feed items popping into view in real time as you continually drill down.

That may not seem like a big deal right now while Social Bookmarking is young and Raw Sugar still doesn't  have a critical mass of users. But look to the future and imagine the potential. Think of how we could use a similar method to progressively filter our news feeds -

World -> Europe -> Ireland -> Dublin -> Finglas

Remember those scenes in The Matrix where only the trained observers could interpret the trickling green glyphs against the black monitor background? Only experience allowed them to see the signal as they progressively filtered out the surrounding noise. As we move to an RSS-everywhere, bogjectified world we'll increasingly require our software to provide a similar means of making sense of the real world, in real time.

Technorati Tags: opml, rss, tags

Comments

[Emanuele Quintarelli]

Hi there. It's a few months I'm talking about the evolution of folksonomies with smart guys like David Weinberger, Frank Smadja, Peter Van Dick.

My starting concern was that actual tag clouds usability is really really low mainly because the increasing size of tags introduced in the system makes difficult for users to extract a mental model out of a cloud.

My suggestions was:
- using faceted tags (meefedia docet)
- using hierarchical tagging (rawsugar docet here)
- using ajax and a wiki approach to build the hierarchy

These strategies can be leveraged at the same time to add a structure to the actual tagging.

The main benefits:
- improving tag clouds browsability and scalabilty
- improving serendipity
- disambiguating tags' meaning

Of course the same approach should be introduced in OPML.

More on my post 'Folksonomies 2.0 - The Chaotic Order'
http://www.infospaces.it/wordpress/topics/information-architecture/74

[Bill Lazar]

James, thanks of course for the kind words. I have a question about the grazing software for you: Using either of the linked application, when I drill down only ten items are visible with no page down or more link to get further in the list. Am I missing something (using FF 1.5)?

[James Corbett]

Bill, that's whole idea of grazing the World *Live* Web (Doc Searls' terminology) or the Feedosphere. The live web is not an archive like the plain old World Wide Web - its what's happening in the here and now. Just as we walk around in the real world and expose our senses to what's happening in the moment, not what happened in the past. Its a new interface for a new web - the live web.

Feed Grazing is about being able to zero into the latest items in a feed flow, grazing them, and moving on. No subscription, no aggregation, just uncommitted quick access.