Mozilla Context Graph: It’s time to bring context back to the web | Hacker News Reader

Mozilla Context Graph: It’s time to bring context back to the web

medium.com

50 pointsst3fan9 comments

This is an interesting thought experiment. It seems Mozilla is trying to lower the barrier to entry for the average, small website compared to the giants in the space right now by providing a tool built into the browser itself.

In my experience, this is an ambitious goal. As humans, we don't like change and, as a whole, won't do it without strong reason. Opening your browser means going to the bigger sites (e.g., Facebook) because they're convenient. The convenience isn't due to only thinking Facebook when thinking of social media, it's about what features Facebook offers me as its user.

If, on the other hand, the goal is more for the audience than the small websites, and it's to expose the average viewer to more variety, why not partner with Stumbleupon, Reddit, or one of many other community link-aggregation platforms?

StumbleUpon meets Facebook News Feed in your browser!

What's Mozilla's goal here? Making a play for ads? The appearance of feature parity with whatever Google (and now Microsoft) puts out?

I see a few attempts by Mozilla at social networks using open web technology. "Hello" was one.

I think the distributed, federated approach such as Friendica (et al.) will be more likely to succeed than this, in large part due to data ownership concerns.

That said, variety in this space is always welcome. We, as the users, can only benefit from it.

This isn't an ad play at all. It's an attempt to build features into the browser that help people find new contextually relevant web sites. Say you're on YouTube watching a video. Where are the outbound links? How do you get to other sites that are embedding or linking to the video to see what they have to say? Today you can't because YouTube wants to keep you in YouTube. But what if your browser could tell you about related sites so you could escape YouTube and find great contextually relevant content out on the Open Web? Wouldn't that be nice? Nothing to do with advertising at all.

Right, so it's a competitor to Facebook Graph [1], which uses a combination of data mining (from Facebook) and special xmlish tags that website operators can put on their page [2].

It sounds like if you opt in to this venture, it will use your browsing habits to mine those relationships; I wonder if it will ingest Facebook's "OpenGraph" tags [2].

[1] https://developers.facebook.com/docs/sharing/opengraph/using...

[2] http://ogp.me/

Where are the people who are asking for help with this? Will people who value having fewer distractions on their screens be opted into this great content?

And what about spam?

Does anyone know what type of technology will be powering this? Is Mozilla building some sort of graph database to make this work?

I though this was hosted at https://github.com/mozilla/contextgraph-service but it's inactive since a month.

You should ask people involved in the #contextgraph channel on irc.mozilla.org (https://client02.chat.mibbit.com/?server=irc.mozilla.org%3A%...)

We are just getting started and most of this is still in the design phase. Technology choices will have to be made.

The two most concrete things right now related to Context Graph are probably Activity Stream and the Fathom project. Fathom is an experimental framework to extract meaning from web pages ad can be used for a wide range of applications.

Production builds of Activity Stream can installed via TestPilot.

https://github.com/mozilla/activity-stream https://github.com/mozilla/fathom https://testpilot.firefox.com

Reply on news.ycombinator.com