Monday, February 07, 2005
Ask Jeeves to buy Bloglines?
I know several regular readers of ZS use Bloglines. So, it may come as news to you that Ask Jeeves is planning to buy Bloglines and integrate it into their search system (expected to go live at some point today). Napsterization has some questions about the search strategy.
What they would have trouble doing is getting all the data, structured, organized and pulled, going back more than say, a month. That's because blog posts fall off the front pages (depending on frequency of blogging and how many posts the blogger displays) and go into archives. If you think about how many kinds of blog software exist, which means many different kinds of data structures for the blog post data, which then it's very difficult to get all the various types of data structured into a single database, just imagine how all the variants of those professionally and homegrown blog publishing systems differ for archival posts. Lots of people customize their archives, as I have in MT and other blogs I participate in with Wordpress, Typepad, etc. Spidering and structuring archives is really tough, tougher than getting the stuff on the tops of blogs right. The point is, a comprehensive database of blogs structured well, going back a couple of years, is really valuable. As is the knowledge of how to put that database together, and run it, along with understanding why this kind of search is very different than those done by Google or Ask Jeeves, whose results don't understand the temporal qualities of blog data, or other aspects that make it different.
What they would have trouble doing is getting all the data, structured, organized and pulled, going back more than say, a month. That's because blog posts fall off the front pages (depending on frequency of blogging and how many posts the blogger displays) and go into archives. If you think about how many kinds of blog software exist, which means many different kinds of data structures for the blog post data, which then it's very difficult to get all the various types of data structured into a single database, just imagine how all the variants of those professionally and homegrown blog publishing systems differ for archival posts. Lots of people customize their archives, as I have in MT and other blogs I participate in with Wordpress, Typepad, etc. Spidering and structuring archives is really tough, tougher than getting the stuff on the tops of blogs right. The point is, a comprehensive database of blogs structured well, going back a couple of years, is really valuable. As is the knowledge of how to put that database together, and run it, along with understanding why this kind of search is very different than those done by Google or Ask Jeeves, whose results don't understand the temporal qualities of blog data, or other aspects that make it different.