Getting Answers | Hacker News Reader

Getting Answers

mikeash.com

56 pointstpaschalis21 comments

In my experience IRC is absolutely the worst place to ask technical questions. You have to be there at the same time as someone who knows the answer, you have to stick around and be ready to respond if someone replies, all the while multi-tasking to filter out all the other conversations in case the responder doesn't use your nick in the response, once you get a response you either have to clunkily weave your conversation into a bunch of other chat or use a direct chat, and the chance of anyone else ever seeing your question and locating the answer is tiny, going to zero if you use direct chat after the initial response. And as the author says, you can't paste more than a line of code without breaking netiquette.

A close second is mailing lists, which the author does mention. I think it's pretty well established that the only way a mailing list can be pleasant to read is if every single participant is uses the same mail client and formats their responses identically, taking great pains to format the exact same way, which has never happened.

There is already a solution which naturally encourages all the good behaviour in this article: Stack Overflow and co.

It really depends in my experience: I think I prefer IRC for more complex/specialized/niche topics.

Quick back-and-forth for details is very efficient compared to e-mail or postings somewhere if asking for details is really needed. Similarly, it feels more appropriate for really short questions.

The waiting isn't so much an issue if you have a stable IRC setup (or the channel has good infrastructure), but can be a problem for people that don't have that. Generally, "please ping me if you reply" works fine, and less active channels can be quite good at directing questions to the right person.

And a big thing is audience: if there's only a few people I think will know a good answer, asking in an IRC channel or on a mailing list they use is more efficient to get them to see it. On Stack Overflow I have the risk of not only having it to word in a way that the few people I want a response understand it, but also so that everyone else doesn't feel like it's a bad question, can only ask questions SO finds acceptable, and maybe the right people don't even look at SO, or don't see the question.

This obviously doesn't apply to questions many people from a larger community can answer.

The guide would have been extremely relevant in the days before Stack Overflow and useful search came around along with the rest of the web friendly internet - the lack of date on the article makes me question when it was originally written, in those days one had to pray someone posted the same question on Usenet or else one would have to go on a road trip to a book store, do an extensive real life document search, then failing that, hope that the technical book store might have something to address the issue or, failing that, copious back issues of IEEE periodicals/Dr Dobbs/etc at the college library (or the ones one had in personally library first).

In my experience, StackOverflow has become oversaturated, while IRC is as good as it ever was.

There's so many questions on StackOverflow these days you're counting on luck that someone (qualified) sees your question. Most questions I've asked in the past year never get to even 10 views, or 1 answer. Another problem is that you're not supposed to respond unless you have The Answer, so if you don't happen to get an expert to read it, you often get nothing at all. Some fields seem to be perpetually full of beginners -- Lisp-related tags are often loaded with "help with hello world" questions.

StackOverflow also has the Close Police, who seem to patrol questions looking for any excuse to close it. Even after 10 years, I can't predict when a question might be flagged.

On IRC, there might be only 5 people actively participating in a channel, but 100 lurkers. When you ask a question, people respond with suggestions even if they don't have The Answer, and this often leads to a workable solution.

The average quality of response on IRC is much higher, as well. A lot of the lurkers are actually semi-famous developers. On more than one occasion, I've asked how to do something like $(app) does, and someone says "$(dev) here wrote $(app), just ask them!".

StackOverflow is from the social media generation of software, where you need a pretty avatar, and accumulate points, and there's rules you must not break. IRC is from the pre-web generation of software, where there are no absolute rules, and you just want to get shit done.

Weird. I was on IRC for many years before Stack Overflow (and crap alternatives like Experts Exchange, Yahoo! Answers, etc) existed, and I have pretty much the exact opposite experience. For example, while SO tags certainly have different distributions of expertise, out of my most recent 30 questions (involving mainly Python, Django, Docker, Rust and WebExtensions) only four have no answers yet.

> StackOverflow is from the social media generation of software, where you need a pretty avatar, and accumulate points, and there's rules you must not break.

Unlike IRC, where people need to learn and adopt the particular culture of a channel, and (in extreme cases) need to supplicate before the right people to get a civil response.

Stack* is about questions and answers, learning and doing.

> Unlike IRC, where people need to learn and adopt the particular culture of a channel, and (in extreme cases) need to supplicate before the right people to get a civil response.

All (software) IRC channels have basically the same culture. It's essentially what you mean when you say StackOverflow "naturally encourages all the good behaviour".

I have no idea what you mean by "supplicate before the right people to get a civil response". I don't think I've ever seen that. Unlike StackOverflow, you have the option to ask one person directly, but you're never obligated to.

So nice article, so poorly titled on HN. Kind of irony here.

Missing tip: when creating a new question on Stack or traditional forum: make your question topic describe the root of the problem. Be concise. You will increase chance that somebody that potentially knows the answer will come to help.

Titling your topic "problem", "need help", "my app doesn't work" isn't going to help you and will frustrate others, usually dealing with people breaking half of the tips from the article.

Sorry for the title, I preferred keeping the original one. Mike's blog has quite interesting stuff!

As for your other point, yes, with all the negativity that StackOverflow sometimes exhibits, I really feel that the pressure/urge for users to create a "Minimal, Complete, Verifiable" example [0] is part of the site's appeal, and part of the reason of the high quality of responses you sometimes find there.

[0] https://stackoverflow.com/help/mcve

Read and understand the output before asking for help. I know this sounds ridiculous or even condescending, but anyone who has worked in this field long enough knows what I mean.

Yeah, output often includes a solution, yet people often just copy-paste it somewhere (or directly to you) and ask what to do.

I include an intranet URL in many error messages at my company, users ignore them and still ask.

That's a great idea (of course, you can't let it rot).

I keep realizing that asking good questions is a truly important skill in any profession and any situation.

Couple of books I found useful :-

A more beautiful question - Warren Berger Power Questions - Andrew Sobel Secrets of Question Based Selling - Thomas Freese.

I feel that what we describe as "soft skills" and that many people discredit, like being professional, being able to clearly communicate problems, solutions, ideas or being always willing-to learn could be better labelled as "Professional Skills".

While they're not a substitute for technical adequacy, at the end of the day, we have real people as co-workers, clients, bosses, and we provide services to them, and those skills are of utmost importance, and should be part of the interviewing process.

Also, thanks for the resources, I'll be checking them out. ;)

As I read this, I ruminated on my travails at getting a standard USB mouse to work on Ubuntu 18.04. I didn’t ask any questions in forums, largely because all of my hours of searches revealed an entitled community responding to other questioners with NIH (It’s a hardware problem, even though your hardware works on other OSs.), instructions that demonstrably don’t work, or just not responding at all. I don’t mean to take anything away from the author’s fine article; it truly is good. But there can be toxic communities. There can be problems that are entirely opaque to the uninitiated. (Who configures mice in 2019?) And there can be problems for which Googling sends you down a maze of dismissive and entitled responses.

Also, if anyone here has gotten an Anker vertical USB mouse to stay awake, scroll smoothly, and click reliably on a clean install of 18.04, I’m all ears. :D

I guess it's not what you want to hear, but that probably is a hardware problem. It may work on other OSes, but if it's not behaving correctly, then it's probably not a well-behaved USB device. And there may be software workarounds--but it's probably ultimately a hardware problem. So you should probably complain to the manufacturer and tell them to get with it and either fix the mouse or upstream a Linux driver that works around its bugs.

Mildly unrelated, but there should be a nice and polite in-person variation of RTFM. It is certainly annoying when certain questions are asked but they are just a google away.

Indeed. But the worst thing someone can do is leave a comment like that on Stack Overflow. Invariably, SO shows up at the top of google's search, so later people with the same question will read that and say "duh- that's how I got here."

My go-to line on this:

> If you spend 5 minutes Googling this you’ll know more than me.

It’s short enough, if flatters and encourages the recipient, it demonstrates modesty.

Much better than RTFM!

I disagree that ESR's HTAQTSW is condescending. The fact that people, over years of time, have written and updated that guide for the benefit of others is the opposite of condescending. IMO calling it condescending is tantamount to violating rule #7 from this guide.

And I disagree that "RTFM" is a useless answer. When it's correct, it's the most useful answer one can get. To put it another way, it's only useless if one refuses to accept it--which is, again, like violating rule #7.

There might be a place for an abridged version of HTAQTSW, but I don't think this is it, because it doesn't improve upon it, and it denigrates it.

While somewhat related to point 8 (Follow up after getting an answer), my pet peeve is people who post about a problem they are having, not getting any response, then a final "Oh, never mind, I figured it out".

Auuggghhhh! The forums you use are a great resource not just for you but for others. A posted solution could help other people who are searching for the same question.

Reply on news.ycombinator.com