Hacker News Reader

Comment by CGamesPlay | Hacker News Reader

What sorts of questions do you ask that lead you to come off as superior? How have you changed your approach to avoid putting off your students?

View on HN

I think that the issue isn't the questions themselves, it's the interaction that results. Student: "why is my Java application crashing", "That's a stack trace. Does it say anything useful?", Student: "No I don't see anything.", It looks like we aren't seeing the whole trace. Have you scrolled up?" etc... the more this goes on, the more it feels quite awkward even if the result is positive. Maybe I'm bias having been on the receiving end in the past.

I've changed my approach by not letting things get out of hand. If I notice that the process is taking a long time or someone gets frustrated, I either let them know about the process or just take a different approach, like just answering the question. The challenge is that, at least in the research setting, I often don't know the answer either. So if the student thinks I know the answer, and I'm just asking them questions, it comes off poorly. So I'll let them know I don't know the answer either, and that we are working through it.

I'd love to see an alt site called SocraticOverflow, where people answer everyones programming questions like this :)

I know a lot of the time people on stackoverflow only have a problem because they haven't asked themselves the right question.

Of course, there's still a lot where it's just a gap in knowledge that needs filling in.

This is interesting. Let's think it through.

For the teachers and students, it would need to be realtime. A teacher couldn't write out a string of follow-up questions in advance and hope the reader answers them in exactly the way you intend.

For the observers, you need them to be engaged with the content as well. I am imaging a sort of dialog tree would be useful here, where a new reader could add an alternative answer to a given question which would then be a candidate for other teachers to add follow-up questions. How would [1] look like if it were resolved using the Socratic method?

For programming in particular, building out an interface to encourage experimentation would be critical. Take for example [2], a recent question from StackOverflow. The questions I would ask to lead the student to the answer are: - What is the actual difference between iPhone and desktop for this code? - What nodes are being selected by the call to `$`? - What makes you confident that the call to `$` is even happening? - etc. The student's ability to even answer these follow-up questions will make the socratic method painful when conducted through a text-primary medium.

This raises another point that teachers need to already know the solution, or at the least have much better guesses than the students (and those guesses need to be generally correct for the site to provide value). This greatly limits the number of people who would be capable of being teachers on the site.

[1] https://stackoverflow.com/questions/10059594/a-simple-explan...

[2] https://stackoverflow.com/questions/46570102/jquery-on-not-w...

I second, @charlieflowers endorsement below.

This is a brilliant idea and it would like to see it.

If the Socratic method get stuck, ask a different question.

Can questions so implicitly contain the answer that one can practically make a statement with a question ?

that is such a brilliant idea. I'd love to see it.

There's a hidden benefit to taking it too far. If you annoy someone with your constant simple questions, they will finally determine that asking you to solve their problems for them will take too long, and they'll try to solve it themselves. Or they'll ask someone else. Either way it's one less annoying student :)

I have to deploy this method with internal customers. I call it either proof-of-work or back pressure.

Reply on news.ycombinator.com