Language wars in tech are common. Java vs C++, Functional vs OOP. If you're doing functional is your language functional enough?

So to in CS Education. What's the best language for CS0? For CS1? Drag and Drop or text based? Functional? Object Oriented? Compiled? Interpreted?

The battles rage on.

One particular "war" that I've been a part of deals with the idea of a language's being authentic or real. I've been a part of this on two fronts.

I guess most visibly this has been with Scratch and other drag and drop languages. People sometimes perceive me as being anti drag and drop. This is not the case. I love drag and drop languages for the earlier grades. My issue is that I haven't seen an advantage to using them in the later grades. I haven't seen evidence that knowledge transfers when an older student - say high school 10th grader or beyond starts with a Drag and Drop language and then switches to text based. I've also called out what I see as disadvantages of drag and drop languages in the later grades - what I call jigsaw puzzle debugging or programming where students write and debug programs by seeing which block fits geometrically as opposed to thinking through the problems and also the small screens students usually use are a limiting factor when writing complex, large programs.

I never thought that drag and drop languages like Scratch were inauthentic or less real than, say Java.

That said, there are those who do. I think it comes mostly from parents and students who feel "they don't use it in the real world." Probably also from hardcore programmers who, even though scratch is Turing Complete, feel that it's less of a language.

I've never taught scratch in a formal class so I never had to deal with any push back. I've only used it in one offs like teacher professional developments or working with students for some side event or project.

On the other hand, I've heard the same complaints when teaching languages like Racket (nee Scheme) and NetLogo. They're not real languages. Nobody uses them in the real world. I'd particularly hear it about NetLogo which is a wonderful language, particularly for visual modelling and simulation.

Never mind that languages don't have to be used in the real world in order to be real and to be of tremendous value in the classroom. Racket is wonderful for teaching functional programming. It's also a super small language with a simple consistent syntax which makes it, when taught right, beginner friendly. NetLogo is interactive, visual and allows for students with little experience to explore complex concepts. In my classes at Stuy, we wrote a maze solver that used implemented a cellular automaton and also explored predator prey models all during one unit in our CS0 class.

Even if a lesser known language, something seen (incorrectly) as inauthentic can truly have an impact.

I saw that just the other day.

I found out that one of my former students was on the team that just won the Nobel Prize in Chemistry. Just seeing the news made my week.

I of course sent a congratulations. As part of his response, he wrote:

You know, I can trace a lot of my interest in computational science back to playing with NetLogo turtles simulating things, exploring cellular automata rules, and graphics.

All the stuff we did in Stuyvesant's required intro CS class.

I don't know if the intro course as a whole made a difference in his life and career - he might have gone down the CS path regardless. It is clear though that using NetLogo in the intro class did ignite a spark. Had we used say Java or C++ from the get go. maybe this never happens.

The whole line about inauthentic languages is really hogwash anyway. Sure, some teachers use the wrong language the wrong way at the wrong time but most languages can be the "right one" under the right circumstances.

In this case, it was NetLogo at Stuyvesant in tenth grade in our required CS class.