Yes, there's complete symmetry between signing and encrypting, thus your proposal is exactly equal to the article's.
The solution is quite simpler. Do not sign text that you didn't create. You create a random timestamped text, encrypt it, and send both to the server.
Or use the Diffie-Hellman key exchange, that as a bonus isn't vulnerable to men-in-the-middle attacks (that both the article's algorithms and the one I just wrote are).
The solution is quite simpler. Do not sign text that you didn't create. You create a random timestamped text, encrypt it, and send both to the server.
Or use the Diffie-Hellman key exchange, that as a bonus isn't vulnerable to men-in-the-middle attacks (that both the article's algorithms and the one I just wrote are).