No, the problem with using lock_keys directly is that it's a run-time solution rather than a declarative solution. In the long run this will make it more difficult to translate to Perl 6. In other words, it's making the same mistake that pseudo hashes did, insofar as it reveals too much of the underlying machinery.