Old News is Good for Google

by Dale Dougherty

Tara Calishain uses the net as a tool for market research. As a consequence, she tracks what's happening among search engines and related services on her site, ResearchBuzz. She has relied heavily on Usenet newsgroups, and the search interface at Deja.com, aka DejaNews. When she learned that Google had bought Deja,
this expert user had this to say:

Even though I speculated on this back in October ("Deja News Search Engine for Sale: News and Irresponsible Speculation," I'm still stunned to hear that Google has purchased Deja.com, including, they say, the entire Usenet archive dating from 1995.


A beta version of the news search is available at groups.google.com , which you can also get to at http://www.deja.com.


Longtime Deja users (like moi) will be annoyed to see that the advanced search suffers from some shortcomings. You can no longer search by title, apparently. You may not confine yourself to a certain group of groups (forsale, standard, complete, etc.) It is not clear if you can exclude newsgroups (for example, do a search for resume and exclude all jobs newsgroups.) Google also doesn't support hierarchical browsing at the moment either (keep repeating to yourself, "It's only a beta. It's only a beta...")


If you're a Google user the search results will look familiar. The search results consist of title and a few words from the message. The bottom line of the search result contains the newsgroup to which it was posted, date of posting, and name of poster. There's also a link to review the entire thread.


The current archive only goes back to August of 2000, but Google co-founder Sergey Brin told me they hope to get up the archives from 1995 within the next 90 days. He also noted that there's some functionality currently missing (the incompleteness of the advanced search, for example, and the lack of Web-based posting) but that Google does intend to restore the data archives to all their glory. "Twenty years from now this will be very valuable data. We want to keep it safe and make it available."


Google does plan to integrate text-based advertising into the Deja searches, both with the premium spots (the colored boxes at the top of the page) and the AdWords program (the text ads in small boxes down the right side of the page) but they didn't have current specifics available. (Searches for general keywords cars and flowers found no advertisements, so I would guess it's not implemented yet.)


I'm relieved to see that Deja is getting what looks like a good home. I can't wait for the promised improvements in the search interface.



It is good news that these Usenet archives appear to be in the hands of a company who understands their value.