In "Does running an alternative full node implementation benefit the Bitcoin network?", I asked about running an alternative full node implementation for the benefit of the Bitcoin network.
What is the current status of different alternative full node implementations with regard to
- maturity (alpha/beta/stable)
- performance
- compatibility tests
- code review
- suitability for different purposes:
- experimentation
- merchants
- as a wallet
- as a lightweight wallet back-end
- solely for the benefit of the network?
Update: Moved the list of implementations into an answer.