unbehagen

I loved XBMC and still use Kodi to this day. Back then, I even proposed and POCed what's now a part of their Add-On system, essentially a fuse-like virtual file system forwarded to Python. Before that, each Add-On had to bring its own UI. This was basically my first OSS contribution and the community was really supportive and welcoming.

Amazing! Very similar approach, would love to heae what you think: https://github.com/gerkensm/vaporvibe

"As a totalitarian society, the Soviet Union valued eavesdropping and thus developed ingenious methods to accomplish it." (page 2)

If it is a chrome extension and is installed via the Chrome Web Store, it can be updated silently in the background if I'm not mistaken. So in theory, wouldn't it be possible to serve Google with a NSL and force them to silently push a modified update to a targeted user that reveals the private key?

Here's the technical background: ClickSend consists of an Android App and a Java server. Both register with a Google App Engine backend. If you send a file to the Android App, the Java server opens an HTTP server, generates a random URL for the file and sends this as a push notification (GCM) to your Android device, along with all IP adresses of the server. The client is triggered by the notification and tries to reach the server on all IP adresses and opens the URL. This allows for direct streaming of multimedia content (player App needed) or downloading. The data never passes through our servers, so it only works if server and client are connected to the same network. We are currently preparing to launch this so all feedback is appreciated.

No comment yet from anyone at Google after nearly 24 hours. This probably breaks every App using the builtin authentication. If anyone knows a workaround, I would appreciate hearing about it!

oh, I just found out that another user tried the same thing and wrote about it before me: http://news.ycombinator.com/item?id=3768030

This approach to tamper with recently viewed items on Amazon embeds the articles as hidden images:

Of course, this only works if your browser accepts third-party cookies and is inspired by this original post: http://news.ycombinator.com/item?id=3766544 .

Are the results any different from the original? Does using an image instead of an iframe have any effects on the results? What are your thoughts on this?

Well, this is not surprising. If you put less content directly on the first page, users have to click on links in order to view the content that is directly accessible in the desktop version, which generates more page views.