I would claim there is a significant portion of population for whom empirical teaching of CS would be far more kind and efficient than going through the Djisktraist pure theory route. Even Djisktra began with writing actual programs until he grogged how to think about it and then moved to pure theory.