This is more of a statement than a question.

A few days ago, I went through the answer to the question: Why can rm remove read-only files?

I found that the accepted answer to it was incorrect. The answer has since been corrected after I commented the mistake in it.
The problem was that I already got like around 60 vote-ups and the answer had like 25k views.

People (in this question they were around 60) who have no knowledge of answer are voting up. The people who are learning are getting wrong knowledge (in this case there were around 25k views).

So what does this say about our culture on Stack Exchange? The trustworthiness of Stack Exchange? Stack Exchange has got a good reputation already. Whatever people will read on it (especially the accepted answer), they will take it to be true (well, in this case they have got only misinformation).

Also, what should be done to prevent this in future (from Stack Exchange's side and not the users')?

  • 12
    I'm not sure what we could possibly do to forever prevent all wrong information; a crisis of faith over a slight error in a single answer seems excessive. The system worked great -- you found the error, commented about it, and it was fixed. A shame it took so long, but the fundamentals of the answer (the permissions of the file don't matter, it's the directory that matters) were right, so that's probably what a lot of people cared about. Nobody actually tried unsetting the directory's execute bit, it had nothing to do with the original question Commented Sep 22, 2015 at 14:44
  • Afaik it doesn't happen too often. Others had similar problems on the SO.
    – peterh
    Commented Sep 23, 2015 at 17:46

3 Answers 3


It speaks nothing of the culture. We're safe.

You have to be aware that you are browsing an Internet - and while answers are generally accepted and encouraged by Community, these might still be lacking, wrong or misleading.

You can't help it.

However, you can always edit the question - so if you've spotted the mistake, correct it, improving the answer. The point of SE is to provide Q&A format of popular questions(and less popular too), so that whenever you encounter a problem, you can find compact solution, and solution is kind of public - you can edit it to improve, provide additional insight, or update, if such need arises.

Yet you specify the very root cause of the problem - people who are looking for an answer are usually people, who don't know the answer! Therefore, they might not see the mistake(especially this one, which could be considered relatively small, depending on context).

Note that this is inevitable in every source of knowledge - if you read academic papers, there are often some errors in there - some books contain wrong information too - blog posts, YouTube videos - mistakes are everywhere.

So we're good. The answer wasn't providing the shell script that wiped out your disc, it most likely couldn't affect anyone in a harmful way - so the worst that could happen is that the answer didn't work. And while it's sad to see upvoted and popular answer turn out to be wrong, in the end of the day we focus on our mission - improve it, and continue providing quality content to the users.

PS. Also note that as the Michael said, the system worked - and it worked well, actually. Compare this to the scandal I've recently learned about, where some physicist falsified data to provide (false) evidence to his theory. Now his work lasted for a bit longer than a question, he also received few awards for his work, until it was discovered to be a fraud. As I said - you can never be sure - but we are fortunate to have a community where providing false answers isn't profitable for both sides.

  • Well said. I think the key concept is to remember, as you said, that "it's the Internet". Stack-Exchange does not exist to be a perfect record of correct Q&A style information sharing. Well, that may be the goal but I don't think anyone really thinks it is an achievable goal...perfection. But I think we do a fairly decent job of raising the bar to an appropriate level.
    – 0xSheepdog
    Commented Sep 22, 2015 at 21:20
  • 2
    @0xSheepdog Afaik a group of high-repu users could exist coordinating their efforts to solve the similar, extraordinary cases of highly upvoted false informations. 5 3k+ users acting cooperatively can close/reopen any question, around the same number of 10k-20k+ users can do the same with delete. Their efforts could be organized easily in a chatroom. Btw, sometimes similar little circles are going mad and start to close & destroy important aspects of a site. And if they have enough support from the mods/powerusers, nothing can stop them. It is the other side of the danger.
    – peterh
    Commented Sep 23, 2015 at 17:53
  • 1
    @MatthewRock I don't understand you, it is completely voluntary work. People are cooperating here not for money, but for common goals: for example, to close such highly problematic content as you named. Making voting circles and similar misuses are forbidden, but an open cooperation for positive goals isn't.
    – peterh
    Commented Sep 23, 2015 at 18:39
  • @MatthewRock For example, there was a "delete voter" chatroom on the SO to solve very similar problems as yours. Maybe it exists even now. Or it is common phenomenon on different meta sites, that the community initiates and executes a major re-tagging project.
    – peterh
    Commented Sep 23, 2015 at 18:43
  • @peterh What "mine" problem are you referring to? Anyway, what I mean is that your comment implied creating 5k people commando, hunting for answers that are wrong - this is both very hard, very tedious and not that interesting. Answering questions, editing; someone would have to cooperate this - managing 5k people sure is hard. I want to show you that it's all easier said than done - and it's unnecessary thing to do, since we have the means to protect us from bad content, which may sometimes work with delay, but this doensn't change the fact that they work. Commented Sep 23, 2015 at 19:44
  • Also, this could be not effective if bad content would be provided maliciously - group of privileged users might edit posts after being check, and it could be unnoticed. But as I said - there really aren't many such people(if any) out there, so building a tank to run over an annoying ant is overkill. Commented Sep 23, 2015 at 19:47
  • 2
    @MatthewRock five 3k users, not 5*3k = 15k users. :D
    – muru
    Commented Sep 24, 2015 at 5:37
  • @MatthewRock Yes, sometimes such cooperatives can significantly worsen a site experience. Major changes require the support of the powerusers/mods of the site. About your other not-understanding: 3000 reputation is needed to cast close and reopen votes, and 5 close/reopen votes can close/reopen a question. Thus, a cooperative having at least 5 members of at least 3000 reputation points, can close or reopen any question. Most site is effectively ruled by such cooperatives. Sometimes the process is perverted a little bit, for example if a such group
    – peterh
    Commented Sep 24, 2015 at 6:24
  • @MatthewRock ...goes mad and starts to destruct the site content and harass the new users. In this cases, they tend to narrow the meaning of the "community" to themselfes. (On Unix SE I didn't experience this phenomenon.) In other cases there is a good-willing group initiative which is destroyed by the mods. Contrary the intentions of the founders of the SE, it is a highly politized in the background.
    – peterh
    Commented Sep 24, 2015 at 6:29
  • 1
    Wow, how did I read that five 3k users as 5.3k users again? Anyway, I agree with the narrowing community problem - I noticed this on SO - hence why I am not as much of a fan as I could be. People on SE too often downvote/vote to close simple questions or questions that are asked by obvious beginners, because they aren't asking the question with 300+ upvotes. It's saddening to see people ask something on SE and receive nothing but downvotes and useless angry comments instead - happened to me once or twice too, I had to remind them that they are there to help, not downvote. Commented Sep 24, 2015 at 8:12
  • @MatthewRock Here is the list of the acquirable privileges by reputation, maybe it will help. Yes, this is also a reason, why would be highly useful to part the functionality of the SO into a redundant network. For example, if the Unix SE would allow Unix/Posix programming questions (which isn't the case, at least not de facto), it had the side-effect that Unix programming questions could have another chance if they got a malignant treatment on the SO.
    – peterh
    Commented Sep 24, 2015 at 10:48

First of all, obviously, wrong answers can and do get upvoted. Votes only indicate people's opinion, they do not magically make something correct. Sometimes, many people get it wrong.

However, in this particular case the original answer was not wrong. At worst, you can say it was incomplete. The first version of the answer stated:

All rm needs is write permission on the parent directory. The permissions of the file itself are irrelevant.

This is true, but incomplete. Deleting a file does, indeed, also require execute permissions on the parent directory. However, the main point here is that the permissions of the file itself are irrelevant and that you need write permissions on the parent. The question was asking why a file with read-only permissions can be removed. The answer quite correctly explained that i) the file's permissions are irrelevant and ii) that you need write permissions on the parent.

Finally, and most importantly, the answer was edited by the OP five minutes after posting to include a quote that explicitly mentions the execute permissions:

Execute permission is needed on a directory to be able to cd into it (that is, to make some directory your current working directory).

So, the answer was actually correct for all but its first five minutes of existence. Yes, the OP's text didn't explicitly mention execute permissions but the quoted text did. I don't really see that there was a problem here. Everything worked as it should and the people who upvoted did so because the answer was fine, despite the omission at the top.

  • 1
    (1) You seem to have a pronoun problem.  The OP (author) of this (meta)question is not the author of the answer that this (meta)question is talking about.  In fact, the author of that answer (ire_and_curses) doesn’t seem to be active in this meta-thread at all.  So why are you using second-person pronouns (“you” and “your”)?  (2) The sort of person who would ask a question like “Why can rm remove read-only files?” quite possibly wouldn’t understand the relevance of the paragraphs saying “Execute permission is needed on a directory to be able to …”.  … (Cont’d) Commented Sep 23, 2015 at 19:43
  • 2
    (Cont’d) …  After all, I can type rm foo/bar without making foo my current working directory, and there’s no reason to expect that anybody who’s reading that question will know what an inode is or what it signifies.  (3) That said, I agree with your main point — the answer was incomplete and slightly misleading.  It was far from “100% wrong”.  (4) Obligatory xkcd: “Duty Calls”. Commented Sep 23, 2015 at 19:43
  • @Scott I do, indeed. I was confused by the phrase "I got 60 upvotes" in this question, thanks.
    – terdon Mod
    Commented Sep 23, 2015 at 20:13

There are types of information if one looks at it :-

a. No information

b. An answer that could be a right answer or not depending on the context it is.

c. An answer which is right for that moment in time.

d. A quick and dirty work-around.

e. An answer which would be more expansive and actually frames in with different contexts of time.

Now if we had a choice, I would always like to go with e. and if not that not c. and so on and so forth. But for e. to happen both the answerer should have enough time to give sufficient depth and at the same time he should be amply rewarded for taking time to answer the question from various contexts and perspectives. But more often than not, I have seen people upvoting d. rather than e. or c. but I cannot blame the community for that. They upvote the solution which works for them. You will always see that it is only a tiny community which is interested to know the genesis and the different reasons, perspectives of state of affairs (whether it is a simple tool or an mini-ecosystem like systemd or something else altogether)

I also do not believe that all actors are trying to mislead people intentionally as well simply because I don't think you can benefit in longer-term because of it. You may get some short-term gains but then that is nullified if you indulge in that behavior.

The idea of stackexchange is quite interesting. You have game mechanics (in the form of points), you have community and you have knowledge-base both in the form of wiki as well as interesting questions that people ask.

While we should strive to improve, I do not think we are lessening the quality or should improve the quality of the site at the cost of not being accessible for millions of users.

The good thing from your question is do we have enough incentive for people to curate both the question and answer and I do not have enough knowledge of that to say it in one way or the other.

I hope the answer sheds a bit more light on what you are asking.

You must log in to answer this question.

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