All resources have DELETE available through HTTP. Any client can make that request and the server responds appropriately. This client makes DELETE available on ‘self’ links. The server may have implemented an OPTIONS approach but this case we haven’t implemented this on the current client.
Please note that APIs that implement batch delete are likely to break these rules—we may look to implement ‘batching’ from a user perspective because that is a good user experience but still implement it as per those rules in the client implementation. What is that rule about types of optimisation as the root of problems?