7

There are many questions asked when stack first begun becoming popular, which deal with somewhat "fundamental" features of a popular language, library or development environment.

The problem with any popular X, however, is that they tend to change.

While 10 year old questions about Fortran may still hold up quite well to scrutiny, questions about languages like C++ or environments like browsers, do not.

I'm sure many of you have ran upon many such example.

For the sake of discussion, here's one I stumbled upon recently.

How to center absolutely positioned element in div?

This is quite a "relevant" question. Centering an absolute positioned div is a task many of use could come up against at one point if we are doing web development. It's also relevant due to the implications it would bare for anyone new to CSS, as they might take away from this how absolute element's behave.

The problem here being that it was answered 8 and a half years ago, when a critical property (transform) was either not part of CSS or not widely spread through browsers.

Meaning the top 2 answers are essentially wrong. Well, not quite wrong, but worse, misleading. They give "broken" solutions which would hurt potential viewer's understanding of CSS, besides provide an ugly solution.

This is quite an issue, because the question has 1M views.

Obviously it's a reader's responsibility to be wise about these things. But, not being able to rely upon top answers to provide a correct solution is a huge drawback when using stack. Especially when we are dealing with the kind of questions someone that's been programming for 1 week might ask. It both makes the experience more annoying and promotes bad behavior among new developers.

In light of this, should there be some way to "necro" popular threads and re-open them for a team of moderators to pick the correct answer every few years ?

"Team of moderators" here could mean anything, an easy solution would be people with some reputation in the area of the question and validation from the community.

However, the way this should be done is not what I want to get stuck on here, but rather if it should be done at all.

I feel like this will become a bigger and bigger problem in the next few years as stack gets to be more widely used and questions get older and older.

  • 2
    Please note that the tags "moderator tools" and "moderators" specifically relate to elected/appointed moderators, not users with sufficient reputation to moderate content. The sort of thing you're asking for is specifically not in the realm of responsibility of these diamond moderators. Only the person who asked the question can accept an answer on a post, no one else. Addressing out-of-date questions is absolutely an important issue that has been discussed before: meta.stackexchange.com/questions/191789/… – Catija Aug 15 '18 at 0:50
  • 4
    4 years ago with 145 views seems to be hardly "important" thought. Is re-opening a discussion not worth it ? I couldn't find any more recent posts. – George Aug 15 '18 at 0:58
  • Also, note, I'm not asking "How should users help with outdated questions" but rather "Should specific answers be added to deal with outdated questions" – George Aug 15 '18 at 0:59
  • 4
    We don't "close" discussions unless they're somehow a bad fit for the site, either being off topic or too broad or having other issues. Also, since you've tagged this "stackoverflow", you might find better results on that site's meta, for example: meta.stackoverflow.com/questions/272570/… It's certainly an issue other places but it's a bigger issue on SO. The very first recommendation on the answer I linked in the first comment is "add a new answer", so even if it's not an identical question, the answer tells you what to do. :) – Catija Aug 15 '18 at 1:01
  • There's nothing wrong with asking here, I'm not saying it's a bad question... here's another more recent example of one here. All of my comments here have been designed to clarify a few errors in what you've either written in the question or tags or comments. – Catija Aug 15 '18 at 1:17
  • 1
    @Catija: If "moderator tools" refers only to diamond mod tools, what's up with the name of the 10k privilege? (I dislike this meme of "only diamonds should be called moderators", and I don't think it has historical support in SE's design.) – Nathan Tuggy Aug 15 '18 at 5:32
  • 1
  • @NathanTuggy I'm specifically talking about the tags I removed. The tag guidance for moderator-tools specifically limits the use of that tag to diamond mod tools. The current term for the 10k tools is just that, 10k-tools. – Catija Aug 15 '18 at 11:41
  • @Catija: Ah, OK, I didn't get that you were talking about the tags. – Nathan Tuggy Aug 15 '18 at 20:05
  • @NathanTuggy Ah. OK. :) Edited the comment for better clarity. – Catija Aug 15 '18 at 20:41
1

A comment already suggested how to deal with new answers to old questions by referencing What is the best practice 'etiquette' to tell an answer is outdated and should be updated? and How do we encourage edits to obsolete/out of date answers?

In that case, it's best to post a new answer that gives new information.

Alternate solutions (such as adding an "obsolete answer" flag to some answers in some fashion) should be pursued through actual feature requests, even if discussions are a good place to start.


A corollary to this is if you have a question that seems like a duplicate of an "outdated" question—but the older answer cannot be answered in a meaningful way today because its nature makes current answers incompatible. (In other words, you can't just post a new answer to the old question.)

In that situation, I would propose the following:

  1. Create a new question.
  2. Reference the old question.
  3. Mention that the old question is no longer relevant because of new technology.
  4. Ask how the question should be answered today.

If you already have an answer, you can also post that to the new question yourself.


Although I see where you're coming from with wanting to remove (or de-emphasize) answers that were correct at the time but are now "misleading," I don't think that anybody who is familiar enough with the questions and answers they're looking at would normally be misled by something that seems "antiquated," especially if they look at the dates involved.

I recognize that this is an ongoing debate, but I'm not convinced that it's actually such a dire situation that there's a systemic flood of people causing harm because of making use of information that's many years out of date. I would think that if things get close to that point, somebody actually will file a feature request—and enough people will be concerned by it that action will be taken.

Are you aware of a specific outdated question that actually is doing harm to a significant number of people because of its existence? If so, I would certainly raise it here (or in the meta of its site) as a cause of particular concern.

I know you do bring up a specific outdated question. But can it be shown to really be harmful as it stands?

In general, as with votes on questions and answers, things on sites like this seem to happen organically. When enough people think a certain way, this is reflected in how things happen. Still, this is a good starting point. Or, if going by previous discussions, a good continuing point.

You must log in to answer this question.

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