With some luck it'd attract enough developers to keep the project alive. Or maybe somebody would figure out a way to monetize the effort. Either option, it'd be much better than just let the service die.

It's not unprecedented. Google did exactly this with Google Wave.

Sadly, I don't think this will ever happen. This project is too small to justify the resources needed to clean up the code and solve any dependencies. And, differently than Chromium / Android / Google Wave, Google Reader is not a platform / protocol / OS. It's just yet-another pet project for a company that - sooner or later - has to justify its investment to Wall Street.

This isn't the first, and certainly won't be the last, product that Google creates just to later destroy it, when it discover it has no business or economic justification.

Keep this in mind when developing for Mirror API, or anything else that is not core for the company.