David Duey

Just Call Me Defrag Dave

A sure sign of acute dorkism:  I'm pee-in-my-pants excited that I'll be going to the Defrag conference in Denver next month.  Can't wait!  Can't wait!  Can't wait! Can't wait! ...

Viral Knowledge

If you’ve tried Geni, you understand why it’s so viral; it’s quick, fun and easy to use and you can send email invitations to family members as you create the nodes of your network.  Someone dubbed Geni as the “most viral application in the world.”

I was wondering if there’s any possibility of creating a viral knowledge graph using the same kind of paradigm that social networks use to grow their membership roster.  In my quest, I came across a slideshow (pdf) from the Moscow Aviation Technical University titled “Viral Knowledge Acquisition Through Social Networks.” (it’s only 10 slides, but it’s a slow download).  The slideshow only presents enough detail to whet your appetite, but there’s enough “meat” to provide a general idea of what they’re thinking about.

The cool thing about the knowledge graph (for lake a better term) is that it’s very, very dynamic.  Social networks are dynamic, but taken to the limit, once the social network is complete, it then just has to be maintained.  Births, marriages and divorces all have to be accounted for, but I’m pretty sure that the dynamic nature of social networks doesn’t even approach the dynamic nature of the knowledge graph.  I'd bet that knowelege is created at a much higher rate than newborn babies.

Digg and Reddit may be good examples of social propagation of information (and possibly knowledge), but I wonder if there’s not a better way to propagate, proliferate and create knowledge than our current blogs and wikis.  Freebase is certainly a start, but it’s not really viral.  Could a modified Geni model work for knowledge?

Social Networks -> Knowledge Networks

Peter Rip posted a really juicy entry about social networks evolving into knowledge networks (well, there's actually more to it than that) titled, "More Than Who You Know, It's What You Know."  It's smells of the stuff of the implcit web. 

Ratings and Reviews

Heather and I have been working away on the alpha release of Georneys.  The sole focus of the alpha version is on the “lost child” problem; to provide a means to have lost children returned to their guardian(s) quickly and safely.

However, as part of the Georneys “product mix” (service mix?) we’d also like to help families answer the common weekend question, “Where should we go today?”  We’ve found that parents often have a difficult time finding new and fun things to do with their young children.  Families typically do the same things over-and-over again; go to the same park, mall, and restaurant.  I think a lot of the repetitive behavior is because familiar things are a known quantity, and with small children, a familiar routine is often welcome.  But, I think a lot of the “sticking with the familiar” behavior is because it’s often difficult to think of new places to go and there’s little time to research and think about it, so the familiar is the default. 

As I’ve given more thought to the “Where to go, what to do?” question, I've seen a number of opportunities to use the internet to help solve the problem.  It’s really an interesting problem and I think most of the current approaches are a bit weak.  What I see very often is that places and events are visually presented on a map (Google Maps, Microsoft Earth and the like).  I think that’s probably a mistake.  If I’m looking for fun places to go, I’m concerned about the distance of the place and the time it takes to get there, but I don’t really care about its location on a map – until I need directions to get there.  I think maps are very useful, but when my question is, “Where is a fun place to take the family today that within an hour and a half driving distance that’s will fit my budget?” I don’t really want a map.  What I want is really good, targeted suggestions of places and events that the whole family can enjoy.  Reviews are good too, but I don’t (and probably can’t) spend my day reading reviews; I want a good summary.

The other issue I have with current services is that user reviews may not always be the best indicator of “what’s good.”  Actual behavior is probably a better indicator; I think it’s somewhat analogous to the attention issue with web content.  Behavior is important, so I think feeding a “where to go/what to do” system with actual data would be very useful.  Fred Wilson talks about some of these issues in his post titled “Taste Neighbors.”  It’s a great post and points to a lot of good opportunities in the user generated ratings and reviews space.

Needless to say, I’m excited about working on some of these problems.

The Implicit Web and Social Network APIs

I’ve been thinking a bit about the implicit web and social network APIs.  I can see were companies like Lijit (I still need to install the Lijit wijit) and Confluence Commons can have a huge impact in the future.  As social network APIs become available, like the Facebook API, the APIs can provide a lot of information that can be used to enable the implicit web. 

The initial services that companies like Lijit offer may not be nearly as important as what they can potentially offer in the future.