"The Data Dump: Fun with Graphs and Charts" 
Marc Hedlund David Hornik; Ian Kallen; Eric Lunt; Roger Magoulas; Adam Messinger; 
David L. Sifry


-----------------------------
contrubutors:
Gabe Hollombe - gabe@avantbard.com - http://avantbard.com
-----------------------------

Dave Sifry
- state of the blogosphere report
- blogosphere doubles in size every 5.5 months over the last 36 months
- over 60x larger than 3 yrs ago    
- over 100k new blogs each day, aprox 1.5 blogs per sec
- 50% of new bloggers are still posting 3 months later
- 10% of blogs update at least weekly (3M blogs)
- lots of spam; about 9% of new blogs are spam; 60% of pings are from known spam 
sources; technorati blocks spam pings before the register as splogs in their 
indexes
- 1.2 million legit posts/day
- news cycle that's shifting.  not 24hr cycle like old days, but one that's 
measured in MHz. (cute)
- blogs brought us a "frictionlessness" of publishing capabilities
- power law: MSN gets the top end, but blogs start to dominate the middle and low 
end
- blogging is big in japan (started taking off about 4 months ago)
    * japanese posts tend to be shorter, but they post more often
- almost 1/2 of all posts use tags/categories
    * about 24% use the rel="tag" microformat

Eric Lunt (feedburner CTO)
- going to show Feedstorm (animation)
- yes, there is a growth in feedburner feeds over the last 2 years
- demo viz created by manifestdigital

Adam Messinger (Gauntlet Systems)
- building a nextgen source control feature - no more broken builds (builds on each 
checkin)
- Lotka's law (80/20 rule) another statement of power law, theorizes applies to OSS 
projects, that few developers write the most code
- going to reply about 2dozen opensource project checkin histories
- ActiveMQ checkins by author follows Lotka
- Lucene checkins by auth doesn't fit as well
- Hibernate EJB3 shows exponential scaling
- conclusion: true that a few devs do most of the dev work 
- more data avail at www.gauntletsystems.com

Andy Edmons (Windows Live Search) 
- showing viz of popular queries by date (rose bowl near jan 1, for example)
- the attention of the world is mirrored in search activity
- hot spots around the globe (often searched for in MS Earth; "walmart" bigger than 
"america")

Roger Magoulas (O'Reilly)
- need to figure out how to compare things when scales vary differently (tracking 
trends)
- 79% of AJAX job postings are in the bay area

Ian Kallen (technorati)
- tracking blog spam
- financial incentives amplify abuse (duh)
- some posts are not easy to id as spam, but all the links tend to point to the 
same site
- lots of splogs use the same name server and belong to same ARIN IP block







== Resources ==


__END__