Sunday, March 31, 2013

Open Source Project Mimics Yahoo Pipes on Your Own Machine

Web applications are so very convenient. But there’s always a rub.
When you move your life onto the web, you lose control of your data. And — who knows? — your web app of choice may even vanish from the face of the Earth while you’re still using it. Think Google Reader.
The growing Indie Web Movement hopes to alleviate these problems, working to give web users more control over their online data and identity. The latest example is an open source software tool called Huginn.
Huginn works a lot like popular web services such as IFTTT (“If This, Then That”), Zapier, and Yahoo Pipes. It’s a platform for running web bots — called “agents” — that can gather information from across the net and ping you when certain things happen.
The difference is that it’s open source software that runs on your own server. That means you control the data it collects and creates.
Huginn can notify you when your favorite sites are updated, aggregate your location check-ins over time, or warn you the day before it’s supposed to snow. But the most interesting tool may be the one that monitors keywords on Twitter and notifies you when everyone starts talking about stuff you’re interested in.

But that’s just the beginning. Developers can build more “agents” for the platform. The software is named after the ravens of Norse mythology, Huginn and Muninn, who travel the world collecting information for the god Odin.
But the point, says Andrew Cantino, the creator of Huginn, is that it’s not Pipes or IFTTT. “Pipes isn’t really being maintained,” says Andrew Cantino, the creator of Huginn. “If This Then That seems to be doing well, but I don’t really know if it’s going to stick around. Ultimately, I don’t control the data.”
Cantino says he created the software for his own use, and he almost didn’t open source Huginn. But he’s glad it did. The response has been enormous. It garnered over 1,200 stars on it first day on the source-code-hosting-site GitHub.
Cantino doesn’t have any plans to turn Huginn into a business as of yet. Before seeing the response to the application, he wasn’t even considering the idea. He works for the web-based project management software company Mavenlink, which doesn’t leave him with a lot of time on his hands.
“If I did turn it into a business, I would have to be very careful about how I did it. People need to own their data,” he says. He’s reluctant to create a hosted version of the application, saying some of the tools — such as site scraping — could be easily abused. But if he did, he would make it possible for users to export their data.
In the meantime, he wants to make it easier for users to get started hosting their own version of the software. He says that once it’s deployed, Huginn is pretty easy to use. Unless you wan to create custom agents, you don’t need to do any programming. But it would be hard for a non-program to set the application up to begin with. He says developers on GitHub are already working on a version that is easy to deploy to Amazon EC2 or Heroku.
He also hopes to see the number of agents grow. He thinks some sort of directory of agents will prove useful, and perhaps a marketplace for people to sell agents they’ve built will make sense.
Nothing is perfect. The sites that Huginn monitors can change or disappear. But it does give users a permanent repository for the data it collects. It’s a welcome addition to other indie web projects like decentralized wiki inspired by GitHub and created by Ward Cunningham — the creator of the original wiki — and ThinkUp, a tool that lets users backup and analyze everything they post to social media sites.

No comments:

Post a Comment