oreilly.comSafari Books Online.Conferences.
Articles Radar Books  

Feed Your Head at Etech 2003

by Daniel H. Steinberg

Editor's Note -- As the Emerging Technology Conference came to a conclusion on Friday, April 25, 2003, many of the participants were still just as enthusiastic as they had been on day one, and the sessions were as full as ever. Daniel Steinberg was there from the beginning, and he's pulled together some of his favorite observations for this article.

A Feast for the Mind

Friday morning, attendees at ETech have that same look they wear when dessert is announced after a huge Thanksgiving dinner. They've already consumed so much. They just want to sit quietly with a cup of coffee and digest everything while talking to friends about this shared feast. On the other hand, just like the pumpkin and apple pies, the remaining sessions and events do look good. Perhaps just a taste. Which one? A little of both, please. Whipped cream? Oh, sure.

Lunch went beyond merely satisfying physical hunger; it was also a time to feed the mind. [Photos by Derrick Story.]

And like Thanksgiving dinner, the ETech attendees are busy packing up the leftovers. They are exchanging cards and email addresses, downloading presentations, and beginning conversations. From hardware hackers, to sociologists, to developers, to journalists, to--you name it, this has been a gathering designed to promote synergies. Tim O'Reilly and Rael Dornfest followed the voice that whispered in their ear to build it and trusted they will come. This isn't heaven, it's Santa Clara. Here are some of the leftovers I've packed.

Sociologist Speaks to Geeks

Marc Smith is a sociologist trying to understand the social aspects of online communities. Smith works in Microsoft Research's Community Technologies Group studying social cyberspaces. He's trying to answer questions about places where groups meet for collective action that spontaneously emerge. In particular he's interested in these collections that are more than the sum of their parts. That links back to the Tuesday morning keynotes of Bonabeau and Rheingold. As these ideas recur and reinforce each other themes emerge at this conference that is more than the sum of its parts.

Smith does not have an easy task. You can walk into a room and make a quick and fairly accurate assessment of the current state of the room. Sure, you don't know what happened before you got there, but others know that about you. When you enter a location on the Web such as a newsgroup or a chat room, you don't have many tools for determining what's going on. Smith is working to change that experience.

If you try to get information on a newsgroup you can get a list of previous posts that include the subject, the sender's login, the date and time the message was posted and its size. Smith notes that this developer created summary of a newsgroup "is an interface to a social space defined by the most antisocial people on earth."

He asks the audience how they might determine which of the Buffy the vampire slayer groups would be the best one to join? As a first pass, he presents an interface that allows you to enter a fragment of a newsgroup name. You will then see a list of newsgroups together with statistics that include the numbers of posts, posters, returnees, lines per post, replies, repliers, and unreplied to posts. This information can be useful. For example if four hundred of the fifteen hundred posters were active the previous month, you get a feel for the size of the core group. Smith says that generally the core group is about 2 percent of the size of the group and they generate more than 60 percent of the content.

Similar statistics for individuals tell you something about them. For example, some authors can be classified as "answer people". Look at the percent of posts that are replies. A high ratio of reply to posts may indicate an answer person. How many different threads is the author posting to. A high thread to post ratio again helps identify an answer person. An audience member raises the question that these might also be characteristics of spammers. Smith agrees but says that other metrics help sniff out spammers. For example, someone posting to more than one hundred newsgroups is unlikely to be a real person. He acknowledges that there will be false positives in identifying answer people but wants to take advantage of the mostly correct information that can be divined from newsgroup usage.

Other metrics are being used to visualize the activity of newsgroups using a tree map. The tree provides insights into the structure of message threads. You can see how many days a particular person has been active. Smith says that 67 percent of the people post only once. Again a small minority of 2 percent post more than three days a month. You can display information for inter-newsgroup or intra-newsgroup.

"The key to a healthy community," Smith says, "is the ability to get other people to come back, be engaged, and interact." Microsoft has identified metrics for measuring the health of a community. To measure the retention of leaders they look at the average number of days that the top forty authors are active. Interaction is healthy when the percent of posts that are replies is high and when the percent of posts that have no reply is low. Size and growth can be measured by the number of posts and the percent change, but Smith asks whether continued growth is always desirable. You can also measure the topical focus by looking at the percent crossposted. The speed of the time to reply is also crucial and can be measured by the number of questions closed and the average time to closure.

You can check out the early stages of this work at http://netscan.research.microsoft.com. You won't see some of the more visual tools yet. Smith shows the audience visualizations of newsgroups. Recency is indicated by color, the number of posts is indicated by posts. The vertical position in the window corresponds to the number of days active and the horizontal axis corresponds to the number of posts per thread. The result of this visual mapping is striking and immediate. You can get a feel for a community and determine whether or not you will feel at home there. Other visual tools include one that indicates on a per author basis the number of threads the user posted to that were or were not initiated by the author. The Netscan data can be used by people to evaluate themselves, others, and their newsgroups.

Pages: 1, 2

Next Pagearrow

P2P Weblogs

Richard Koman Richard Koman's Weblog
Supreme Court Decides Unanimously Against Grokster
Updating as we go. Supremes have ruled 9-0 in favor of the studios in MGM v Grokster. But does the decision have wider import? Is it a death knell for tech? It's starting to look like the answer is no. (Jun 27, 2005)

> More from O'Reilly Developer Weblogs

More Weblogs
FolderShare remote computer search: better privacy than Google Desktop? [Sid Steward]

Data Condoms: Solutions for Private, Remote Search Indexes [Sid Steward]

Behold! Google the darknet/p2p search engine! [Sid Steward]

Open Source & The Fallacy Of Composition [Spencer Critchley]