Interview with Gabriel Weinberg (DuckDuckGo)

As everybody knows, the world of search is dominated by a sole brand: Google. Yahoo!, Bing and a few others try to follow, but the greatest part of market is basically in the hands of Mountain View - with several shades along with the lights. One above all: the fact that users are tracked, and the resulting problem of the filter bubble.

Anyway, there are many other alternatives which could boost well the variety of information retrieval: why not to try them? So we have interviewed Gabriel Weinberg, the young creator of one of our favorites' - DuckDuckGo.

Ok, let's start with the nasty one. There are loads of search engines on the web: why should one pick up yours?

You should use DuckDuckGo as your primary search engine for way more instant answers, way less spam and real privacy. We put zero-click info above the results in a red box that often answers your search without clicking. We try to more aggressively ban all those useless sites with just ads. And we do not track you.

I find very nice the presence of many goodies. Tech goodies in particular. Can users suggest more of this?

Absolutely! That's how a lot of them originate, actually. Got any ideas :) ?

How does the "!bang" keyword work?

You can add an exclamation point and then keyword to the end of your search and then it will automatically search another site for you. For example, bags !amazon will search for bags at Amazon. There are about 1000 keywords so most big sites just work, and then there are shortcuts, e.g. !a for amazon, !yt for youtube and !w for wikipedia.

Technically, how do you cut off all the clutter and the all-ad site?

We are constantly crawling the Web and looking for those "parked domains." We then maintain a list of them and make sure they do not appear in the result set.

You provide also "instant answers" to popular questions. Where do you find the data for that? (I assume the sources must be very reliable).

Yes, we deep-crawl a list of trusted sources, and in some cases send them your search on the fly. The biggest deep crawl source is Wikipedia and the biggest on-the-fly source is WolframAlpha, but there are about 50 sources now and growing. The premise is that for any given search there is usually a vertical search engine out there that answers it better than a general search engine. What we're trying to do is get you that information from that right site so you don't have to know which site to go to (and you don't have to actually go there).

Which are the best ways to improve one's ranking?

The best way (for all search engines really) is to get links from high quality sources, e.g. newspapers, educational institutions, etc. Not an easy task, but it works!

[26/07/2011]