-16

Disclaimer: This is basically a re-post of a meta question I posted over at Programmers.SE

TL;DR - 404 is a Bad Thing (tm). When StackExchange increases 404 errors, it's Making The Internet a Worse Place. (PLUS: I'll add from Bart's comment: "Deletion of questions seems to lead to quite a few dead links. Is there anything we can do about that?")

Prelude: I first posted this on Programmers, because I first noticed it there, but thanks to Yannis' very helpful and complete answer there I would now guess that this is a StackExchange wide problem.

So here's what I wrote:

I have been hitting quite a few 404's (...) lately. I was wondering how great the idea of perma deleting questions (...) really is.

(...)

Not saying that these questions are a good fit to the site, but removing them completely, leaving dangling links all over the internet (am I exaggerating? :-) certainly isn't my idea of Making The Internet a Better Place.

Note that I don't mind closing; I wouldn't even mind a trash.stackexchange.com; (...)

And, while I'm at it, I take the liberty to include a small part of the most upvoted answer by Stargazer. (Well, I guess it may qualify as a rant, but hey ...)

Are you making the web a WORSE place?

This one has an answer:

YES!!!

If the goal of software developers is to help the user with meaningful error messages, then I can confidently say that you are making the web a worse place. You're creating 404 messages despite the fact that the intent of the user was obvious.

Don't say that no one notices. This thread exists because people noticed. You're making the web a worse place.

Improve this question
11
  • 14
    It took quite a bit of reading (between the lines) to see that you seem to be asking "Deletion of questions seems to lead to quite a few dead links. Is there anything we can do about that?". Would that be a fair summary?
    – Bart
    Sep 27, 2012 at 19:28
  • @Bart - Good summary, yes.
    – Martin
    Sep 27, 2012 at 19:41
  • 3
    You might want to make that your tl;dr. I have a hunch most of the downvotes are because at first glance this seems like a rant. But that's just my guess.
    – Bart
    Sep 27, 2012 at 19:42
  • 5
    So what can we add here that isn't already adequately covered by Yannis' excellent answer on the original question you cited?
    – user102937
    Sep 27, 2012 at 19:43
  • 6
    "I take the liberty to include a small part of the most upvoted answer by Stargazer. (Well, I guess it may qualify as a rant, but hey ...)" What's notable is that you took the least important part of that answer. The answer was talking about the nature of the error message. His problem is that 404 technically means "user screwed up" (though virtually no user actually knows that), when it should be a different error that signals that the content is not available.
    – Nicol Bolas
    Sep 27, 2012 at 19:44
  • 1
    @Bart Actually I think the majority of downvotes are from meta users who are tired of hearing about deleted questions and having to defend deleting them so much.
    – Rachel
    Sep 27, 2012 at 19:45
  • 4
    @Rachel Or, in my case, it's a result of understanding that 404 is exactly the result that must be returned for deleted content, and disagreeing with any suggestion otherwise.
    – Andrew Barber
    Sep 27, 2012 at 19:46
  • @Rachel True, downvotes might still happen based on the content. But less so if it doesn't sound like a rant from the very start.
    – Bart
    Sep 27, 2012 at 19:47
  • 2
    @BenBrocka Actually that probably is the view point for non-moderators because we can't search deleted questions. The majority of times we come across deleted questions is from links, and typically the deleted question contains some useful content to the user clicking the link. Nobody links to truly bad questions that almost all users would agree deserves to be deleted, so we never see those.
    – Rachel
    Sep 27, 2012 at 20:05
  • @Rachel - "Nobody links to truly bad questions that almost all users would agree deserves to be deleted." - Look at the answers the diamonds gave. They seem to disagree. They Know Better. Shrug
    – Martin
    Sep 27, 2012 at 20:07
  • 3
    lolwut? . . . .
    – user102937
    Sep 27, 2012 at 20:34

5 Answers 5

Reset to default
19

Returning a 404 causes the search engines to remove deleted posts from search results, which is what we want. So you can't categorically say that 404's are bad. Eliminating 404's prevents us from taking out the trash.

People posting links on their blogs or other websites are responsible for policing their own broken links, whether they originate from SE or somewhere else. This is how the Interwebz have always worked.

Improve this answer
35
  • 5
    Sooooo you're calling the many highly voted, very useful questions and answers that have been linked to elsewhere on the internet many times and since then been deleted for not meeting SE's question standards "trash"?
    – Rachel
    Sep 27, 2012 at 19:37
  • 3
    You can send a 404 header and still have the content, just sayin.
    – user159834
    Sep 27, 2012 at 19:37
  • 17
    @Rachel: If they've been closed and deleted, that's exactly what they are.
    – user102937
    Sep 27, 2012 at 19:38
  • 2
    @WesleyMurch: Why would you want to do that? You're saying the resource doesn't exist, and yet it does.
    – user102937
    Sep 27, 2012 at 19:38
  • 7
    @WesleyMurch: There's nothing to preserve. Deletion occurs because the material was deemed unworthy of preservation.
    – user102937
    Sep 27, 2012 at 19:39
  • 4
    @WesleyMurch It's a technical/semantic issue. We don't want the content, and we don't want anyone to think we have the content. 404 is exactly the result to return, as a result.
    – Andrew Barber
    Sep 27, 2012 at 19:44
  • 2
    @Robert: "If they've been closed and deleted, that's exactly what they are." Unless, you know, the mob is wrong about what should and shouldn't be deleted. I've never understood this fervent belief that moderators hold that everything is where it ought to be. A deleted question is bad by virtual of being deleted. Just like a person accused of a crime is guilty because he's accused, right? Carts and horses work best when done in the traditional order.
    – Nicol Bolas
    Sep 27, 2012 at 19:47
  • 3
    @BilltheLizard A deleted:1 search link? No fair, these only work for diamonds ;)
    – yannis
    Sep 27, 2012 at 19:50
  • 3
    @NicolBolas: Nobody argues that genuinely bad questions should be deleted, should stay deleted and should thereafter never show up on anyone's radar. Almost everyone (with the possible exception of Rachel) universally agrees that popularity is not a good metric for evaluating question quality. There have been many requests for some "limbo" where these popular questions can be safely held for viewing by the terminally curious; they are always declined. In short, we've been there, done that.
    – user102937
    Sep 27, 2012 at 19:53
  • 5
    @Rachel When I put my trash cans out every week, I don't mind that some random guy comes by and picks up the soda cans and scrap metal from my trash. I certainly will if he says that I should not throw them out and that I should instead collect them in my front yard just so he can pick one whenever he wants. But see, he's a real nice guy and doesn't demand that. He understands that this is my trash, but has some value to him and so makes it his responsibility to ensure that he gets the most out of it. This is where I don't get all you SE crap huggers. Why not just host them elsewhere?
    – Lorem Ipsum
    Sep 27, 2012 at 20:07
  • 3
    @yoda That's not a good analogy since SO is a community-run site. Suppose a couple had a big collection of photos from their life. Periodically one or the other (or sometimes both) would put the ones they consider "bad" in a separate box. Its not a problem until one of them decides to dump the whole box in the trash and get rid of them, because the value of a photo from one person is not the same as its value to another. So although one person might see a photo of a kid making a goofy face as trash, the other treasures that photo and has a valid complaint that it is now GONE.
    – Rachel
    Sep 27, 2012 at 20:21
  • 3
    @Robert: No, the inevitable conclusion to my logic is, "I don't trust the system, and therefore we need to make the system more trustworthy". Notice how I never said anything about what should be preserved or is visible or whatever. My contention is that your fervent belief that "deleted post=crap" is wrong. That's what I'm arguing.
    – Nicol Bolas
    Sep 27, 2012 at 20:29
  • 3
    @Robert: Yes, and I don't agree that the safeguards are "adequate". So I find this blind faith you have in the accuracy of the system to be very troubling, especially since you are part of the system. When the powerful start thinking that they're perfect and have perfect judgement (or "adequate safeguards"), that's a problem. I'm not saying that the system itself is bad; I'm saying that having such faith in the system is bad.
    – Nicol Bolas
    Sep 27, 2012 at 20:35
  • 3
    @NicolBolas: Then post a feature request asking for the changes you believe will make the system "trustworthy" and the safeguards "adequate," and let's see what happens.
    – user102937
    Sep 27, 2012 at 20:36
  • 3
    @Nicol I'm sorry, but that's not really the argument. If it was, then you'd be better off fighting it with your undelete votes than commenting here. If 3 people vote to delete something (which is after it was closed) and then no one votes to undelete it, then you can't really say that there has been an argument on whether it should be deleted or not. A lot of the questions that were culled early this year had between 10 and 20 delete votes. If it's deleted by a moderator, then bring it to meta and argue why it should be reopened. I think Shog had a long post outlining what needs to be done
    – Lorem Ipsum
    Sep 27, 2012 at 20:51
6

Most posts are deleted with a reason. The most obvious is that they do not fit in the image of a SE site.

If you don't delete these posts, people see these and are confused about the true image of the site and don't know what to ask or answer here. But if you delete them and keep the 'true' posts, the overal confusion is reduced. Making a better site (and a better web) in the long run.

If it wasn't for the Google ranks, it would have been a nice idea to point to a page stating that the content has been removed because it was not suitable for the site, with a link to the FAQ and the front page.

Improve this answer
8
  • 4
    I'm pretty sure that this approach would penalize us in the Google rankings. Returning a 404 is the semantically correct approach. This is what happens when anyone anywhere on the Internet removes a page or other resource.
    – user102937
    Sep 27, 2012 at 19:36
  • Valid point. I will change the text.
    – Toon Krijthe
    Sep 27, 2012 at 19:38
  • 3
    Actually, this is exactly what SE does. When you browse to a deleted page, there is a 404 page that states "This question has been deleted for reasons of moderation," with tips on how to find similar posts.
    – user102937
    Sep 27, 2012 at 19:48
  • @RobertHarvey The problem is the "404" takes up half the page while the "this question has been deleted for reasons of moderation" is written in small regular sized font. The majority of our users associate "404" with "Page Not Found", and even the title of the page says "Page not Found"
    – Rachel
    Sep 27, 2012 at 20:09
  • 1
    @Rachel You're already here, post a feature request about it ;)
    – yannis
    Sep 27, 2012 at 20:10
  • @YannisRizos I'm busy, I need to go home and play some Borderlands 2 :p Maybe tomorrow if nobody else posts it
    – Rachel
    Sep 27, 2012 at 20:11
  • @Rachel, You have time to play games... scary. ;-)
    – Toon Krijthe
    Sep 27, 2012 at 21:29
  • 1
    @GamecatisToonKrijthe lol why on earth would that be scary?? :)
    – Rachel
    Sep 28, 2012 at 11:30
5

"Deletion of questions seems to lead to quite a few dead links. Is there anything we can do about that?"

What's the alternative? We have deletion as a feature of the site. It's a vital tool for enforcing community standards about what should and should not be on this site.

If a post is deleted, then any links to it need to not serve that content. Otherwise, there's not much point to deletion. We can debate about what kind of HTTP error we get, but that's a pedantic and irrelevant issue. The point is, the user is not going to get what they expected.

There's no "solution" to this. We have historical locks for questions that we deem are interesting and important enough to keep around. But that's about it. If it's not one of those and it gets deleted, then you're going to get some kind of error when you access it.

We can talk about what kind of error you get. But the link will still be dead, no matter what.

Improve this answer
12
  • 2
    I'd rather get a notice that "Content has been removed" instead of "Page not found". To me, Page Not Found means a bad link. Content Removed means it wasn't any fault of mine or the person linking it, but rather that the site hosting the page has removed it.
    – Rachel
    Sep 27, 2012 at 19:54
  • @Rachel: Fair enough, but a dead link is a dead link. Also, there are technical issues around HTTP errors that may make such a solution unworkable.
    – Nicol Bolas
    Sep 27, 2012 at 19:57
  • 3
    @Rachel Hm? The current page says: "This question was removed from <site name> for reasons of moderation. Please refer to the FAQ for possible explanations why a question might be removed." and it even has links to similar (not deleted) questions. There's a 404 image there, but the text clearly explains that it's not the visitors fault the content isn't there.
    – yannis
    Sep 27, 2012 at 19:58
  • 1
    @Rachel: Take a look at this page
    – user102937
    Sep 27, 2012 at 19:59
  • 1
    @YannisRizos I never noticed that before, I think was blinded by the huge "404" image that takes up half the screen and the mental association between "404" and "Page Not Found" that comes from being a programmer. Perhaps change the 404 image to something else (and the "Page Not Found" title)? (I'm testing on Programmers.SE since my profile there is the fastest way I know of to find a deleted question)
    – Rachel
    Sep 27, 2012 at 19:59
  • @Rachel Log out (or set your browser in incognito mode) and then take a look at the page Robert linked to (since you have 10K on SO, and can see deleted content).
    – yannis
    Sep 27, 2012 at 20:00
  • 1
    "This question was removed from <site name> for reasons of moderation." -- this is a NEW error message that was not there when I posted the meta question over at P.SE. I may be wrong and blind, but would someone point out when this feature was introduced?
    – Martin
    Sep 27, 2012 at 20:00
  • @Rachel Changing (or removing) the image, and perhaps changing the title of the page would be an interesting feature request.
    – yannis
    Sep 27, 2012 at 20:01
  • @Martin This has been there for as long as I remember...
    – yannis
    Sep 27, 2012 at 20:02
  • @Yannis - As I said, I may be wrong and blind. Really, I never noticed this before.
    – Martin
    Sep 27, 2012 at 20:03
  • @Martin Or my memory may be failing me (I blame tequila)
    – yannis
    Sep 27, 2012 at 20:04
  • 2
    No, that's been in place for a couple of years at least.
    – mmyers
    Sep 27, 2012 at 20:07
2

There are two things I would like to say:

  1. This 'problem' exists and will exist on all sites that accept user content, simply because not all the content worth keeping or the *user wants it deleted

  2. In most of the cases, if the content is trash, it will not be linked, therefore the amount of broken links should be low.

* This is not the case in SE sites, the license you agree basically grants them the rights to keep posted content against your will (however, you can have your name disassociated)

Improve this answer
2

Also posted on the original meta question at P.SE:

What if we returned a 404 error message, but the custom error page showed the original content of the page (at least until the page isn't getting hit from search engines any more)? This solves the problem with the user interaction, because I've run into 404's for programming issues on SO myself, and it's always just a little irritating when I'm in a rush.

Showing the original content helps the user, even if the original content had a big "closed as dupe" message, linking to another question. Proceeding with the 404 helps search engines avoid sending people to that link, and will help clear the way for more relevant content to be linked to.

After the page no longer appears in search engines (a few weeks after deleted?, not sure time to propagate), stop displaying the original content, and go to a standard dead link page, which will force other sites with outdated links to update. Other sites should have had ample time to see that their external link was returning a 404 and update it at that point.

Improve this answer
1
  • It's a good idea, but this still requires two modes, as some posts (having spam or offensive content) must be actually put into the trash and forgotten.
    – user102937
    Sep 27, 2012 at 23:04

You must log in to answer this question.

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