Duck Duck Go

[mparks]

I have used many many search engines but I was reading up on the new Linux mint 12 and found that they are integrating https://duckduckgo.com/ into the new version, so I took it for a spin and gave it a try. I must say duck duck go is really neat. It pulls in a lot of info from many sources, have a look for yourself and check out t there Goodies sections
https://duckduckgo.com/goodies.html and they have an open API

Here is there API page: https://duckduckgo.com/api.html

"Our Zero-click Info API gives you free access to much of our topic summaries (API example), categories (API example), disambiguation (API example), !bang redirects (API example), definitions (API example), and more (API example)."

I think we could make use of this.

What do you think?

[jitterjames]

sure.  well as it is we can use it with our scrape command and regex

eg: 

scrape   http://api.duckduckgo.com/?q=new%20york&format=xml&pretty=1

in the next version I will have a new scrape command that makes it easier to pull xml tags from pages like this and rss feeds.  In this case we could pull the <Definition> tag easily enough.

The tricky part would be to make a command that works for the different results that you can get depending on what you are searching for.  It seems inconsistent.  Sometimes you get good stuff back in <Abstract>, other times in <Definition>

[nime5ter]

Update: James recently posted the XML for a new command that makes decent use of DuckDuckGo to provide answers to basic questions.

It uses the RoboBrowser plugin for VoxCommando, some regex magic, and TTS. See: http://voxcommando.com/forum/index.php?topic=789.msg9800#msg9800.

As always with commands that rely on third-party websites, if the website changes radically the command may break (or at least the regex may need to be updated to respond to the changes).

We used this just the other night to impress a guest, who posed the question: "How far away is the sun?"