CS profs are talking about small programs. A sort routine, a program to parse a file, maybe a compiler. These are challenging problems but once you think of a solution the actual act of programming it is fairly easy.

SE profs talk about large systems. Many people working for years on the same application, all doing a small bit. The hard part lies in getting correct requirements and hoping they don't change too soon. Integrating all the little parts into one big working system.

Programming is easy. Software engineering is still an unsolved problem. And, in my opinion, not any closer to a solution than in the seventies.

Compilers are small?

Compilers aspire to be something like a total function. Define you inputs and outputs and you are "forced" into a correct implementation.

real software isn't like that. I interpret "small" to mean localized in idea space.

Depends on the compiler/language.

Most programming _is_ easy, with ‘just’ a lot of details that have to be tracked precisely.

The CS profs likely are a bit like math profs, in that they leave such picky details as an exercise. Strings are abstracted as sequences of characters, not the relative mess of UTF-8, time runs forward every second, months and years, let alone leap years and leap seconds do not exist, time zones, let alone time zones changing over time, do not exist, etc.

The software engineering profs likely know how much time and effort solving those trivial smaller problems take.

I think in theory you just assume optimal implementation. In engineering this is more of a luxury.

Both programming and CS are easy. But I’ve never heard a CS prof admit this in public.

Estimating is hard for sure

Programming is easy. Writing correct programs is hard.