Hacker News Reader

Comment by zengid | Hacker News Reader

>>[Smalltalk is] the perfect instructional language for teaching programming to both children and adults alike. It’s very easy to learn and extremely versatile...

Where's the data to support this? Has anyone done a randomized double-blind study on new programming students using smalltalk vs groups using Scheme, JS, or Java? Why is someones biased opinion enough to justify this claim? We need to hold ourselves to a higher standard than to accept unsupported arguments (meaning they havent been tested empirically) [1].

[1] https://www.youtube.com/watch?v=uEFrE6cgVNY

View on HN

Just curious, have you learned ST ? I went through a good deal of languages and I have to admit, ST surprised me more than expected and mostly positively.

I don't like the smug some STers glow when they talk about REALOO but except that it really has some exceptional qualities (self consistency, humane interface, small size of the basic system).

I haven't learned ST yet but look forward to someday; currently I'm learning a bit of Ruby. I don't doubt that it is an excellent language, but is it actually easier to learn without any prior programming experience, or is that claim it just based on opinions?

I don't know any methodical study, so it may only be opinions. That said, after suffering years of bad language features, you feel weird when one actually delivers on some bold claims. Smalltalk is not even a language, it's a system, a self contained system that is quite small, consistent, yet somehow devoid of ceremony (FP languages are small, but one has to read through the culture, ST stay low in intellectualization). I don't see many of the things that tripped me up as a noob, and lots of stuff that I ended up seeing as valuable (how booleans are defined for instance).

To support your claim, people have been learning BASIC, Oberon, and [on kid side] Scratch very well for a long time. I'd love to see Smalltalk with its OOP testing against those. My guess is structured programming in BASIC or Scratch will come out ahead.

Anecdotal: my 8-year old daughter loves Scratch, and I noticed that what she found the most intuitive was the message passing aspect of it. Assuming my kid is average in this respect, Smalltalk might have an edge in teaching kids how to code.

That makes sense. Kids will have passed messages person to person or place to place. Taking turns drawing on the same area is a bit less common. Doing math together exclusively by writing & erasing numbers on same, physical space would be extremely weird. Might be a better metaphor for shared-state concurrency. That's just what popped into my head thinking of the common examples people start out with like bank transfers.

Reply on news.ycombinator.com