[joe_the_user]

Before computers, far more people did bookkeeping or followed recipes or read sheet music than did abstract mathematics. I think one can reasonably say that lists of instructions that change mutable variables/objects is one of the most natural ways people can operate. Anyone learning a task essentially does this in the real world. It may indeed be terrible when done at a large scale. But there are a lot of people who can naturally understand these.

Immutable data may be a great idea. But is it the first that occurs to anyone? I have an MA in mathematics and the idea of mutable variables never seemed strange in the process of learning to program.

But even more, for a lot of people, even algebra isn't a natural way of thinking. What's natural is a series of instruction about how perform arithmetic. Learning what an abstract variable means is a hurdle for a lot of students.

The thing is that "natural", in this context, just means "what occurs to you first, all things being equal".

You may have a good and correct argument that learning Haskell is the right way to do things, that the costs of learning are more than exceed by the benefits.

But I don't think you can argue people into having functional approaches be the first thing that occurs to them.

[dasil003]

I think it's impossible to disentangle the reality of common programming languages and how they work from what is inherently more "natural". The bottom line is programming is not natural, not in any way shape or form. It's hard to learn imperative programming, it's hard to learn functional programming, and it's really hard to learn object-oriented programming (well). If all languages were functional I don't think that would really be harder than learning BASIC, people would just accept that is what programming is and get on with their lives.

I acknowledge that imperative programming is more analogous to real world tasks, but then as soon as you try to write your first program you run headlong into the fact that programming is nothing like other tasks you have done before. The tricks you have to learn to make imperative programming work are not really any easier than the tricks for other paradigms, and in fact you can mix and match from various paradigms to great effect (see: ruby).

[wtracy]

"...it's really hard to learn object-oriented programming (well)."

You have no idea how desperately I wish more people understood that. I could swear that every introductory programming book produced in the last years was either written by someone who either completely forgot how hard this material was to originally master, or who was just some freak of nature who just intuitively grasped it.

Please, please, please, give your students a solid grounding in procedural (or even functional!) programming before teaching OOP. You can teach them about objects and how to use them, but please stop making people write entire classes before they have at minimum a month of coding exercises behind them! (Universities that teach CS101 in Java, I'm looking at you.)

Anyway, I strongly agree with you and reached for the upvote button, and ended up downvoting you instead. Stupid touchscreens.

[majewsky]

There, I undid your downvote with my upvote.

[joe_the_user]

but then as soon as you try to write your first program you run headlong into the fact that programming is nothing like other tasks you have done before. The tricks you have to learn to make imperative programming work are not really any easier than the tricks for other paradigms.

There you go again, equating good and natural.

It seems like even functional decomposition can be unnatural for people even given the commonness of "one big loop" mega-functions as the solution created by a lot of contractors.

Certainly, the problems of mud ball and simple imperative programming styles is that they get harder as you go along.

Indeed, programming methodologies are more or less unnatural ways to program that allow programmers to be more productive and programs to scale in compensation for that unnaturalness.

Consider that mathematicians among the general public, pre-computers, were actually far more rare than functional programmers among programmers. Mathematics is also an unnatural construction. I grew having a great deal of difficulty with handwriting before computers appeared and I learned to love math shortcuts for the writing they saved me. But I also learned most people are more mentally lazy than physically lazy and on average would rather do more work than learn a new thing.