Consider This

Maybe I'm still too new, but I'm going to put this issue out there:

I'm not good with cars. Say I'm out driving and my car breaks down. I call a tow truck and take it to the mechanic, who I tell I'm willing to pay after agreeing to the terms. I tell the mechanic the transmission may be broken, and that caused the problem. I leave my car with him, and leave for home in a borrowed car.

Later in the week, the mechanic calls me and lets me know the car is fixed, but tells me that the problem was not the transmission, but was electrical, and needed a tuneup.


I have 2 options here:

  1. I can be offended because the mechanic figured I knew nothing about cars, and never checked my transmission.
  2. I can trust the experience I paid for and use my now working car until the next time it malfunctions use the same mechanic, building trust in the process.

How This Applies Here

I'm pretty "experienced" at understanding how Linux works. I know that the major difference is the package manager, the release schedules, and at times configuration files. Also knowing the distribution methodology helps determine the types of users in the userbase helps.

That being said, I answered these 2 questions recently:

  1. “GLIBC” can not be used when I have installed the gcc 5.1.0
  2. C compiler cannot create executables

with negative to no vote at all. My issue here is should I answer the question as posed, i.e. check the transmission when it's not the problem, or should I properly fix the car, so that the customer drives off happy?

In both cases, readers have rewarded me negatively because I chose to make the customer happy instead of answering the question as posed. In the first case, telling the OP to politely reinstall instead of walking him through a side by side installation, as the other answer suggests, puts his broken system in a usable state. In the second situation possibly a new user wants to install WINE to play a game, so should I make the problem worse by telling him to install build-essentials, which all the answer posters knew to do including me (installing a building a package for a relative newcomer can be a somewhat daunting task, and notice the OP said install, not compile...), or should I give him his completed car, and have him install WINE, so he can play his game?

Comments Welcome...

  • 1
    I'd suggest also looking at Gilles' answers, he has a very good inventory of well written A'ers. Don't just read them, look at how they're structured. Also go back to when he first started answering, and pay attention to how his style has evolved. Do that for others here, you'll start to notice that many of us were pretty bad in the beginning but then started to hit our strides. Here's Gilles' 1st answers: unix.stackexchange.com/users/885/…. You can move through them by clicking the paginated links in reverse order.
    – slm Mod
    Commented Jul 18, 2015 at 19:14
  • 1
    I think that's called the XY problem.
    – Anko
    Commented Jul 23, 2015 at 22:43

2 Answers 2


Writing answers that are well received takes practice. Before I started answering questions here I looked at answers that I thought were good and tried to emulate them in my approach until I found what worked for me.

For the 1st answer: “GLIBC” can not be used when I have installed the gcc 5.1.0, I'd restructure it so that the solution is at the top and not the bottom. People are generally more interested getting it fixed and moving on, than they are with why. By stating the problem, you're forcing everyone to read through that first, so that by the time they get to your actual solution, they've already had a negative experience.

For the 2nd answer: C compiler cannot create executables, I disagree with your logic for never compiling from source. NOTE: I didn't d/v it though.

For this answer, you've taken a position of scolding the OP and anyone else who's done this, and then began mentioning things that most people will have to go read through to get what you're talking about, specifically who's "Yaegashi" and what does he have to do with the problem at hand.

There are other problems with this answer, and I don't want to completely slam you on it, since I appreciate anyone that's attempting to do something better.

So when answering a question I generally try and do the following things:

  • Be concise.
  • Put the solutions up at the top; put the reasons why it didn't work or ways to improve it at the bottom.
  • Pull the meat of other resources into your answer, don't make people have to chase 2-3 links of material to get your points.
  • Always try and treat the OP as a student, and you're the teacher attempting to explain material to them.
  • Write your answers for a larger audience, and not just the OP.
  • Always provide real examples. Again, Stack Exchange is all about teaching moments for the OP and any future visitors.

I've generally fallen into the pattern of answering like this:

### Solution ....

### Why ....

### References ....

And lastly try and keep a sense of humor about things. Otherwise the Internet will drive you crazy 8-).

Oh, and compiler Q & A has always notoriously had a very low return on investment in terms of your time vs. rep returned, so there's that too.

  • @scott - thanks. Not sure what happened w/ the bullet list there. I had it but it as a list but it looks like it got munged on the save.
    – slm Mod
    Commented Jul 23, 2015 at 6:14

Lets continue with your transmission/electric problem:

I have a friend which car had a "problem with the transmission", because that was what was failing. The only apparent issue. Well, results that his problem was actually electric (after months of going to mechanics trying to figuring out what was wrong with the transmission), which was only visible in the transmission.

What can we learn of this? That the problem isn't always were it appears to be. That's why, instead of going about that you think where the problem is, when asking the question is more important to provide context relevant (and sometimes irrelevant) to the issue you are facing, which then would leave the experts to pinpoint what's wrong and how to fix it.

The only way I find such actions useful, is when you describe what you think is wrong, why, and any attempt to fix it. That way, when someone suggest you a solution, they have all the information which could screw up their methods or having bad compatibility with them.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .