Physics is centuries old though. Computer science is not. That physics changed so dramatically in the first half of the 20th century is certainly surprising, considering that the field already had a long history behind it.
Precisely because physics is centuries old, you'd expect that the things we teach beginners are settled. We teach them even though we know they're wrong, but there is pedagogical value in teaching it this way. I doubt we'll be teaching quantum gravity to freshmen in 100 years (assuming we'll have it solved by then), they'll still learn classical mechanics first.
