DISQUS

One of our blogs on UMBC's new academic programs on games, animation and interactive media [1] shows "unknown subscribers" for the count. Well, I know it has at least *one*. Is there a threshold on the number of subscribers?

[1] http://gaim.umbc.edu/news/

Great analysis! Put this in the "everything I learned about X, I learned from my dog" bucket:

(Google Search)

The 2006 TREC conference had a blog track focused on opinion extraction. NIST also asked for proposals for subjects for a TREC 2007 blog track and two proposals were made for slog detection, one from NEC and one from UMBC ("Blog Track Open Task: Spam Blog Classification ").

My blog reading is contextual and I'd need three rivers, at least. In the morning as I drink my coffee I catch up on news and political blogs. While I'm at work I focus on work-related technology blogs. Throughout the day, to relieve stress or fill an odd five minute gap, I might dip into random entertainment blogs, like Boing Boing.

We reported on the distribution of blogs in the blogpulse dataset used for the 3rd Annual Workshop on Weblogging Ecosystem: Aggregation, Analysis and Dynamics at WWW 2006.

See http://ebiquity.umbc.edu/paper/html/id/299/Char... for a paper with the data.

Our approach to differentiating echt blogs from splogs and from random feeds was to build a training set and then use it to train an SVM model. the accuracy for the blog/non-blog decision was about 98% and for the blog/splog decision was about 88%.

I think your 500 words and two posts a month constraint is quite reasonable. Generous, even.

I generally like to watch the Keith Obermann show on the US MS/NBC cable channel. It's a news show, I guess. One thing that annoys me about it, though, is that the last half hour seems to mostly be devoted to stories that bounced around on the internet that same day. So I think that the show has a bunch of young 'reporters" whose beat is to sit in a windowless room and surf the web. Talk about outsourcing!

While it's probably not controversial to believe that blogs influence politics and society, it may be hard to prove it objectively. An easier task is to show how blogs can influence other blogs and Web based communities. Akshay Java has been modeling influence in blog communities and has a technical report on it: Modeling the Spread of Influence on the Blogosphere. I think the work can be extended to document the spred of information and ideas from blogs to MSM. That's a bit closer to showing that blogs affect society.

Looking at data from bloglines users, we've concluded that

- The feeds that really matter (FTRMs) are a very small fraction of blogosphere
- FTRMs double each year, not each six months
- Most users follow a modest number of feeds

For details, see http://ebiquity.umbc.edu/blogger/2006/04/18/pre...

One thing that's missing, IMHO, is the ability to register your comments with several IDs. I'd like to have my personal ID, but also define it as part of a group ebiquity ID. We could put code to link the ebiquity group ID comments on our ebiquity group blog.