22 ms·
Your mileage may vary. What worked with one student won't work for everyone else. Part of the responsibility of a mentor is to tailor the approach to one that
by BeetleB 10y ago
Your mileage may vary.
What worked with one student won't work for everyone else.
Part of the responsibility of a mentor is to tailor the approach to one that works for the mentee.
I can say for sure: His approach would fail for many (most?) students. It's great it works for some, but let's not generalize it to everyone. I personally am glad I was not introduced to programming this way.
Python definitely can work better than C/C++.
Using an IDE can definitely work better than Emacs/vim.
His approach is almost the ideal - and the thing with ideals is that they work for only a few. It's a very tough lesson I had to learn. I had a very strong background in many technical topics (physics, programming, mathematics), and was known for it. People would occasionally come to me for guidance and advice.
I failed almost all of them by behaving in a manner similar to the author.
Why?
Because the ideal approach, where you must get all of the fundamentals really well before moving on, doesn't work for most. Heck, even I did not learn everything that way. I felt it was better than the way I learned and I was wrong.
With most people, you have to provide a more measured approach. Teach some basics, and then give them fun stuff. Then go back to some more basics, and then some more fun stuff. Insisting on C/C++, with pointers, on Emacs, delays the "fun stuff" quite a bit to the point where many students feel (incorrectly) "Maybe this isn't for me".
An analogy in mathematics: In some universities (e.g. East Europe), mathematics majors are taught analysis before any calculus. Their calculus course is actually the real analysis course we have in the US. Start with sequences, limits of sequences, convergence, metric spaces, compact sets, Bolzano-Weierstrauss theorem, all of this leading up to limits of continuous functions. And only then do we tackle derivatives and later integrals.
I mean, after all, why should we teach people derivatives if they do not understand metric spaces well, right?
Imagine if all engineers learned calculus this way. Most would drop out.
The optimal education for the brightest tier of folks is suboptimal for the majority. (And perhaps vice versa, although I have doubts on that).