I presume the underpinnings of Reader (search, caching, etc.) are so tied to Google tech that open sourcing it would either be impossible or lead to a neutered version that was useless.