Search results relevant to who I am

Dave Winer has a simple, but very good idea.

“A very simple idea for improving Google search and getting rid of spam problems all in one swoop, and it’s not the first time the idea has appeared here. Let me tell you where my weblog is. Then it knows what my interests are. Give me search results relevant to who I am”

Extend that: And / or, let me tell you where my OPML / Blogroll is. Or my tagcloud, or my wishlist, or my book catalog, or my bookmarks.

Tags: , , , , , , ,

Comments (2)

  1. Eran Globen says:

    The simple approach to search personalization has been tried many times before and never worked very well. Do we really need delicious to create a personalized search based on my bookmarks? Nope. Do we need OPML to track what I’ve been reading? Nope. Just because you have a couple more acronyms doesn’t mean you’re in any way close to solving some of the basic problems with personalized search. Systems that use this type of information fail on the following fronts:

    1. The context is too broad – I don’t subscribe to blogs about everything I’m interested in. What about my current context? I read blogs about a dozen different subjects but right now, I’m interested in just one.

    2. How do you catch interest shift in a timely manner? Say I started researching a new topic, I know nothing about it so there’s nothing in my personal context for you to personalize on (not in a good way, at least) what does your system do then?

    3. What about seasonal shifts in interest? big events, elections, big news stories, etc.

    4. Many many interface problems and issues (feedback, control, etc.)

    I agree, some of those are good sources of information. They might be usable for personalization in one way or another. But there’s a lot of ground to cover between saying here’s my OPML file to an actual solution.