We are a user experience design and software development firm
Hire us to design your site, build your application, serve billions of users and solve real problems.

I've had this thing around for nearly two years and have just not known what to do with it. Google was doing a pretty good job with search results as long as a topic didn't have a profit motive associated with it. On popular, money-making keywords, the top ten are polluted by the SEO consultants. What to do?
One option is to use the Google advanced search to limit the search results to recent content. The SEO dudes haven't colonized this territory yet. But this only gets you so far; sometimes you want all of the results. What else could work?
Well, a mashup between Google and Delicious seemed pretty obvious. But messing with Google search results is a no-no. How to get around this? Simple -- let the user do the reranking using a bookmarklet and JSONP.
Thus Rerank was born. I use it on a daily basis to improve my search results. Everyone I have exposed to it has been pretty jazzed by it. Thus my dilemma: I can't scale the backend right now because it depends on del.icio.us, which throttles calls from a single IP. So, I'd like to show off this idea, but if more than a half dozen people use it at a time, it stops working. So I've been sitting on it for all this time.
My colleague Brian Dillard has been pushing me to demo it nonetheless, so I've made a demo movie of it. You can find it here. Maybe someone else has more luck scaling this idea than I have.
Technorati Tags: ajax, bjax, google, seo, bookmarklet
Hire us to design your site, build your application, serve billions of users and solve real problems.
This is pretty cool. I have also been pondering on how to tweak Google search results. Personally I was focusing on the annoyance of getting duplicate results.
For example I search for something that appears on the PHP documentation, then I end up with multiple copies of the PHP manual. Wouldn’t it be cool if Google would actually figure out that all these duplicate sites just differentiate from each other based on the navigation and maybe an ad banner. I guess they could do this by comparing some other pages on the site. Then instead of multiple entries with the PHP manual, they could just present the content once with a list of available sites next to that.
Another annoyance is duplication of blog entries. Some people just blindly copy, others add commentary. Ideally Google would figure out the original source and again just list the places giving plain copies. But for those people adding commentary, it would be cool if the commentary would be available in a sidebar that makes it clear which paragraphs are being talked about in the given commentary.
These two things should also be somewhat possible to realize with a meta frontent to Google.
Aside from that I have also been pondering how to better integrate social networking into the mix. The issue I see is that people can also try to bomb those. Furthermore its hard to find the people that actually know something in a social network. Just going by number of connections is kind of hard, since most social networks do not make it clear who “begged” to get connected and who is getting constantly “begged” to connect.
Comment by Lukas, Friday, January 18, 2008 @ 1:10 am
Have you thought about making it into a Firefox extension or Greasemonkey script? That way del.icio.us wouldn’t throttle your usage, since the requests would all be coming from the computer running the browser instead of a backend server.
Comment by Jarett, Saturday, January 19, 2008 @ 12:14 am
Yeah, this is also what I thought was the way to go. Extension probably more so than Greasemonkey. Actually has anyone done a combination of the two? The Extension installs a Greasemonkey script for the core logic, so that it is easily editable, but you still get the nice update notification functionality from the Extension (though it should notice updates, drop a copy of the new version in a new file so that it can be manually merged).
Comment by Lukas, Monday, January 21, 2008 @ 3:12 am
Wow, i find this meta layer on top of google super interesting: if you do such firefox extension, i’ll definitely jump in, test and provide feedback.
Comment by Alexandre Plennevaux, Friday, January 25, 2008 @ 2:58 am