intertwingly

By Sam Ruby, May 31, 2002.

Inspired by Jon Udell's recent articles, I've written a forward looking update to Manufactured Serendipity, focusing on new ways to manage the experience.

Social networking in Radiospace 

Jon Udell: The strongest correlation connects Sam Ruby, Peter Drayton, and Gordon Weakliem... I'm sure none of those three would find this result surprising.  On one level I don't find it surprising, on another level, I find it very surprising.

If shown your sample, yes, I would have picked out Peter's and Gordon's as the ones I personally identify the most with.  Jon's next closest match would have been a surprise - Paul Snively.  But looking at his subscriptions, I see many sites that I frequent enough that I should subscribe to them.  There clearly is a correlation here, perhaps even stronger than the index suggests.  Enough so that I am tempted to follow other links from Paul's page with the expectation that I might find other treasures.

I can also confidently state that I would not have had any meaningful interaction with any of these three if it were not for my weblog.  This social network didn't survive the transition to blogspace - it truly didn't previously exist.  It emerged and evolved In directions that I could not have predicted.  I've changed it, and it has changed me.  As concrete examples, sites like bOing bOing and Hack the Planet cover topics that I would not have thought I was very interested in a mere four months ago.

Piers Harding, DJ Adams, and Rael Dornfest provide still more examples of people who initially found an interest in my work and now I find myself increasingly finding an interest in theirs.  I first met Rael at the O'Reilly Open Source conference in 2000.  He presented after me in the same room, and I stuck around.  He presented on something called Meercat and talked about RSS.  Didn't seem very interesting in the time.

Discovery Techniques 

Referers, and the incremental enhancement that linkbacks provide give you specific and targeted feedback which is very valuable.   But to truly make accidental discoveries, you need to get access to steady streams of data which is likely to be - but not known to be - related.

One way to make these discoveries is truly by accident.  Surfing to weblogs that operate in firehose mode are excellent mechanisms for doing this.  But this does not work quite as well for sites that produce high quality content on a less frequent and/or intermittent nature.  Subscriptions to the content, aggregated by readers is a great way to increase your rate of consumption.  So, I was wrong.  RSS is very interesting.  

Metalinks provide another layer of indirection - they identify other blog entries that reference some link that you have also. This is a great way to identify common interests where no a priori relationship exists.

HTML Links can provide a way to connect the dots between HTML files and RSS files.  This can provide the means for auto-discovery.

Finally, google has a facility to find related web pages.

Beyond Backlinks 

I don't merely want a backlink - I want that blog entry to appear in my aggregator.  And not just that entry, but probably the next several too.  While at the moment, this is just wishful thinking at this point.  But it seems obvious to me that blogrolls and subscriptions are vital enough elements in the blogging experience that there should be more tools to help manage these resources.

I should be able to define a few parameters, like the optimal number of entries I would like to see in my aggregator per day, and "the system" would discover other news sources and segregate them into sites that I have to visit and sites that I have to be subscribed to.  Ultimately, subscriptions should decay unless reinforced by additional links, at which point the source can be replaced with another one that is now more relevant.

By doing nothing more than occasionally linking to articles I find interesting and writing down my thoughts, the system would learn about my interests and would be able to suggest further reading.