Microsoft Research preps Usenet analysis tool

Researchers at Microsoft are working on technology that makes it easier to navigate Usenet news groups and could eventually help clear clutter in e-mail inboxes, a Microsoft researcher said Tuesday.

A Microsoft concept called the Community .Net Server takes a new approach to displaying Usenet groups and message threads, making it easier for users to pick a relevant group, cutting out spam and displaying the most active threads first. The concept is for Usenet groups, but can be extended to e-mail, Microsoft Research Sociologist Marc Smith said in a presentation at Microsoft’s Mountain View, Calif., campus.

Usenet is a giant, distributed database of discussion groups, called news groups, that predates the World Wide Web. Postings in the tens of thousands of groups can be read with a Web browser or with reader software such as Microsoft’s Outlook Express. Typically Usenet groups are listed alphabetically in a reader, group information is limited to basics such as the number of messages and message data is limited to message size, poster name and time and date of posting.

“We need tools that will help us better discover and use online communities. The interface to a social space was designed by some of the most antisocial people in the world,” Smith said. “You want to see more metadata about these communities, (such as) how many people are in there and come back.”

Microsoft Research loads news group and message data into a SQL Server database for analysis. Users can get a detailed overview of activity in Usenet groups. This allows a user, for example, to find not just a group on Windows XP with a lot of messages, but one where many postings get replied to, Smith said. A proof of concept interface called Netscan is online on the Microsoft Research Web site.

Furthermore, message analysis before the list is displayed to a Usenet user can make sure that only relevant messages are shown, cutting the spam that is prevalent in newsgroups. Through a personalized homepage a user can be kept up to date on replies to posting and keep track of often watched threads, groups and posters, Smith said. A concept of the personalized homepage is not yet available.

“We want to make engaging with communities easier and friendlier with this interface. The tools (to access Usenet) have not evolved while there is so much to go after,” Smith said. Usenet was very popular before the advent of the Web and Web-based chat boxes. Despite predictions of its death, Usenet is well used by many, for example through Google’s Google Groups and for technical support through Microsoft’s Communities Web site.

When it comes to e-mail, Smith believes the analysis technology can replace the “Outlook Today” screen in the Outlook e-mail client. The “Outlook Tomorrow” screen would display message threads, remind users they have not replied to a person they usually get back to immediately and provide other information on the inbox.

However, Smith has yet to persuade any Microsoft product groups to incorporate the Microsoft Research “social accounting” metrics. The Exchange, MSN Search, Windows and Outlook teams have shown interest and some inside Microsoft are using the tools to track the Microsoft newsgroups, Smith said.