We use these URLs to confirm that Encore can measure well-known instances of Web censorship. We consider these URLs low risk because each top-level domain on this list routinely serves resources across origins. (For example, thumbs-up buttons, embedded twitter feeds, analytics, advertisements, etc.)
We use these URLs to verify that Encore can measure different kinds of Web censorship across devices, platforms, and Web browsers. These URLs are operated by our Web censorship testbed.
We gleaned most of these domains through manual perusal of the "Herdict and partner lists" section of http://herdict.org/lists in March 2014. Herdict asks its users to measure URLs on the same domains; we only measured favicons to reduce resource usage and risk to users. We halted measurement of these URLs out of ethical concern. Measurement should have ceased a few weeks earlier but didn't due to a configuration error.
Results for these domains, while interesting, don't appear in our SIGCOMM 2015 paper out of ethical concern.
During development, we crawled several popular sites to discover whether they hosted URLs amenable to measurement with Encore. For a few weeks, we instructed Encore clients to measure some of them. We halted measurement of most of these URLs to reduce data sparsity and resource consumption.