You know, a very common newbie Lisp question is "why is my list still the same after append".
I bet you no Python beginner asks why their list is different after list.append(3), having expected a new list to be returned and the original one left alone.
Imperative is intuitive; the real world is apparently imperative. Though it can be described by functions in which t is just a parameter, the way those functions play out is that the past values of t are not available.
To take a more level view, the expectations w.r.t. append could be conditioned by prior exposure to other languages. However, any exposure to computers at all reinforces the imperative view. Almost all content is manipulated destructively: documents, settings, you name it.
a = [1, 2, 3]
b = a
b.append(4)
assert a == [1,2,3]
No matter which paradigm we use, there are going to be hidden traps new programmers need to learn. I think there might be fewer roadblocks in an immutable language because you can examine each function in isolation. There's a smaller chance a beginner will screw something up on line 10 which causes a problem on line 150. More moving parts makes it easier to get stuck on a buggy program.On the other-other hand, it is a slower ramp up to doing useful things. It's much easier to read/write to files and draw graphics in an imperative language.
If the immutable language has a single equality operator which makes it impossible to tell whether or not a value has been cloned, you might be safe from this issue.
Deep vs shallow operations is an issue in both imperative and functional programming. In python, you have '==' vs 'is' for different types of equal. You have deep vs shallow copy: https://docs.python.org/3/library/copy.html
I beg to differ. The real world is declarative. But as software developers we mostly suck at creating declarative ways to express problems. The spreadsheet being the one example of an intuitive declarative UI that non-programmers grok easily.
Your average corporate VP of engineering uses a declarative interface to issues charters to his developers: "You need to hit X performance target, Y cost of goods, and Z MTBF. By Date D. Go."
Declarative is the way the world works.
The industry will catch on, eventually.
Perhaps an inviting flower is declarative to a bee, or reactive chemicals are declarative to each other.
Other than that, I'm not so sure.
Programmers tend to favor the declarative approach because it leads to simpler programs, but it's only easier to understand if you are trained in (somewhat advanced) mathematics; most people succeed in avoiding learning any mathematics.
But yes, mutability is surprising. It's something to explicitly teach.
Assuming that many of its audience are self-learning/non-CS-careerists, Python's popularity and adoption rate is also going to be a major draw. Learning Python as an attractive job skill is as much an incentive as the intellectual accomplishment.
I'm teaching Python because... well, because I teach Python, and this is my way of helping people who want to learn.
I also believe that Python is a good first language for many people. Not everyone, but for a lot of people. It's not only relatively easy to learn, but it's practical. Many people want to learn Python because of their jobs.
God forbid that the functional crowd gets their way. Most people are terrible at recursive thinking. That's why those languages have such a reputation for being difficult.
The declarative languages are also a good starting point, but the most used one (SQL or spreadsheets) are somewhat orthogonal to what one tries to do with python.
You need for loops. You need if statements. You can't just hope that they will get recursion.
I often wonder if imperative programming is actually a beginner-friendly abstraction - in particular, because of name rebinding and interior mutability. My university's CS course is highly technical, but also targeted at non-programmers (officially, the lecturers cannot expect any computing experience at all) - and it teaches Haskell, which addresses some of those concerns.