Here's an idea for an open source search engine: If you could make the search index small enough, only the crawler would have to be hosted. The actual search

Here's an idea for an open source search engine: If you could make the search index small enough, only the crawler would have to be hosted. The actual search engine would run on each user's machine and download the diffs to the index regularly, and big brother would never see your queries. This might be a pipe dream, as modern search indexes are probably much larger than the plaintext of the whole of the web. Maybe it could be divided by subject.. I guess that would have the same privacy problem: the server would see which subjects you were requesting. Maybe you could download the whole thing and only decompress locally the subjects and languages you wanted.