I think Thomas Ptacek takes it a necessary step further: "Don't implement crypto." Even if you're using a good algorithm, your implementation of it will have ho

I think Thomas Ptacek takes it a necessary step further: "Don't implement crypto." Even if you're using a good algorithm, your implementation of it will have holes. The trouble is that the normal exploratory programming technique that gets a MVP up and running doesn't find security holes.