Hacker News Reader

> I'd be curious about scenarios like: > > - Practice identifying the semantic bug in a 20 line snippet of code. How effectively would practicing this help a person identify real bugs in actual code?

For a couple of years I've been a high school computer science teacher. Having 50 students making the same errors over and over, made me very good at spotting syntactical errors in other people's code. Furthermore, I was also able to easily identify "beginners' bad patterns", which are basically the inverse of the usual best practices. The downside of this was that some students tried to use me as a human compiler/linter because I was seemingly able to spot the problematic areas in their code in almost no time at all. Of course, at such moments I had been monitoring their progress for some time. Nevertheless, it helped me enormously to support my students in learning to debug their own code because I could have them reflect on their own practices and design choices.

Just an anecdote.

+1 on what you said.

I was a TA in college for intermediate c++ object oriented course for 16 classes approximately over 2 years. (about 500 students). So many basic and core programming practices almost always stick out like a sore thumb to me. I usually graded in a text editor without any IDE support, and towards the end there, got to the point of just glancing through there program I could easily mentally simulate the compile errors and if fixed, the runtime issues/behavior as well.

It definitely help make me a much better programmer and developer. Now working in the industry, I still see a lot of things I can't help but wonder "WHY?!" though technically they are often not wrong, but just unnecessarily complex or convoluted.

> - Essay writing. Suppose for a moment that a human essay could be judged accurately enough to create a 500 word essay trainer. How effective would training on it be to quickly being able to articulate one's thoughts?

This exists. It's called an audience and publishing something every day will get you good real quick.

A lot of famous authors started as journalists and more recently bloggers for exactly this reason.

My college's placement test used some kind of algorithmic essay grader. I found it strange. It checked for typos, vocabulary use, etc. probably.

I would be interested in learning about any algorithm capable of grading an essay. Typos and vocabulary use should only account for a small portion of the grade in my opinion. To me, an essay should be graded based on whether or not it accomplishes it's goal, whether that is to inform or persuade. I've always felt that this is subjective and difficult for even a human to put a real grade on.

I feel like it would be really easy to game an automatic grader. I mean, it can't understand what you're writing, so if you use good vocabulary and complex sentences, it won't know if you're writing nonsense.

It was a very basic "essay", only a few paragraphs if I remember right, answering an easy prompt. The goal was probably only to score technical writing ability while disguising it as a normal essay assignment - it wasn't revealed that what you wrote would be graded by an algorithm until afterwards.

More info about the one used by ETS: https://www.ets.org/erater

Gimpel Software used to run "Bug of the Month" magazine ads showing code with subtle C++ bugs that would be caught by their Flexelint linter. Here is their Bug of the Month archive (2002-2012):

http://www.gimpel.com/html/bugs.htm

>Similarly, I'd be curious about a sunk cost rationality trainer, logical proof trainer,

A strong understanding of cognitive biases, logic and informal fallacies allows one to better reason from first principles. Without this, even an intelligent person will only think derivative thoughts because their truth value will come from their similarity to those of authority figures or what they read in a book.

If one can reason from first principles, one can come up with ideas that haven't been thought before and develop them in a structured way without getting stuck in delusions. It's also important to be able to detect mistakes in one's thinking, especially when there is no one else to ask if you're wrong.

That sounds rather unrealistic. The chance that you're the first one to think thought x is pretty low. Rely on other's work and step beyond it, don't try to ignore it.

> Chess problems with 5 pieces on the board. How helpful would practicing these be to solving problems with 6 pieces?

Another anecdote: I experienced a noticeable improvement in my abilities at Chess by studying and playing Go. I'm still not real good at Chess because I haven't studied it, but I can now beat people I didn't used to be able to beat.

I've had a similar experience where playing Go made be better at living life, and being away from the game and living made me better at Go when I expected to be set back. This probably only applies at kyu levels but I did find it surprising.

Sounds like Go is the Lisp of board games?

"Lisp is worth learning for the profound enlightenment experience you will have when you finally get it; that experience will make you a better programmer for the rest of your days, even if you never actually use Lisp itself a lot."

It will be a similar enlightenment experience with Haskell?

> playing Go made be better at living life

Would you mind talking about that some more, please?

If you want to hear something similar from someone who achieved a very high level in chess, I'd highly recommend reading Josh Waitzkin's The Art of Learning.

A completely arbitrary related concept: I'm much better at running my startup now than before I played StarCraft seriously.

StarCraft made me a more strategic thinker and it's why I think it might not surprise people that Emmett Shear (ceo of twitch) is really good at StarCraft.

I played casually in the early days, but a few years back I wanted to get decent enough to play online regularly, and it changed our company for the better.

Do you micro the workers or focus on the macro game?

Macro

Is it your skill at StarCraft, or could it be that you developed strategies for learning, resilience, etc while trying get better, and those are the actual success factors?

Conceptual stretching is my guess. More strategy than normal for me.

Do you know specifically how your play changed?

As I said, I'm not a strong chess player and don't know much about strategies. I think my reading ability is primarily what improved.

Hey - I have been working on chess game with lesser pieces on board - www.halfchess.com ... I initially thought it would be a great way to teach people/kids chess. If you have any ideas for my little game - please share them with me - navalnovel at gmail com