Darkcoin should be aiming for an unbreakable solution, an unhackable solution, anonymity that cannot be overcome simply by owning a large enough % of coins.
We don't want users to simply feel safe, but to actually be safe. Coinjoin, darksend, and any other future derivative that relies on mixing pools of transactions is inevevitably going to fail the anonymity test, as the signer of the pooled transaction has access to information.
The only true solution is an encrypted packet of information approach, where although the packet is visible on the network, the contents - and ability to open said contents - is only accessible to the recipient (ie, their encrypted wallet's private key).
I don't really agree with this. Encryption isn't the only answer. What you pointed out before (Monero and any other CryptoNote protocol coins) actually achieve extremely high anonymity. As shown in the whitepaper (
https://cryptonote.org/whitepaper.pdf), the ring signatures they use cannot be broken. They have a mathematical proof in the appendix that shows the true signer in the ring signature is impossible to determine with greater than 1/N probability, N being the number of outputs being mixed. Your standards may be high, but this is really, really good for practical purposes and would stand up well to blockchain analysis.
There are some other advantages too, e.g. you mix outputs not transactions so you don't need to participate with others sending the same time as you. For the same reason, arbitrary amounts can be sent without worrying about timing analysis or tracking change.