As much as I agree with the sentiment behind this, I think the implementation is slightly misdirected. Neither page popularity or query popularity are necessar

As much as I agree with the sentiment behind this, I think the implementation is slightly misdirected. Neither page popularity or query popularity are necessarily proportional to domain popularity (eg, *.github.com). Ruling out the most popular domains is therefore, I suspect, neither good or bad in terms of the quality of results it produces on the whole. Sometimes it will produce better results, sometimes it will produce worse, sometimes the same. If a search engine/tool is going to add value, imho, the very difficult problem that it must solve is to improve the quality of results. Unless I'm missing something, I don't see that here (yet).