I've noticed a few "late hits" recently. These are answers coming in months after the accepted answer. In one case of the answer came in 17 months late!

In that case, since the answer was a reference to a blog post written by the answerer, I questioned him in comments as to whether that was appropriate. I felt it was a bit spammy.

In a second case the answer is another tool recommendation. I wouldn't call it spam, since I don't know if the answerer has any association with the tool.

Are these cases something we should be concerned with? There aren't too many of them, so maybe we can just leave them alone.

  • 29
    I will answer this question in 6 months – juan Feb 27 '10 at 22:29
  • Only six months? Why not 17? – John Saunders Feb 28 '10 at 1:33
  • @John, you're hitting the nail on the head here: there's simply no difference between 6 or 17 months. (Or "current" and "late" in general.) – Arjan Feb 28 '10 at 3:48
  • 9
    I'm the one who asked the question in the first place. Thanks to John for submitting his blog post. I enjoyed it and encourage anyone else with useful information on the subject to do the same. When I accepted the first answer, there had only been a few submitted in the first place. As time went on, the answers got more and more informative. I see nothing at all wrong with submitting answers long after the question has been asked and an answer accepted. – Kilhoffer Mar 1 '10 at 13:20
  • 1
    @Kilhoffer, you know you can change the accepted answer, right? Just click the "Accept" tick at another answer. – Arjan Aug 7 '10 at 9:27
  • 4
    Please close this question, as it was asked six months ago. The answer cannot hold, it is too late... – Andrew Grimm Sep 12 '10 at 13:03
  • What? A year has passed and no new comments have been added?? Delete that old crap! – mlvljr Feb 22 '11 at 18:28

12 Answers 12


I find that I object to characterizing such answers at "late". Acceptance may mean that the original author and dealt with the problem in front of them and is moving on but it does not mark any kind of "end" to the question. Any question that is not "too localized" remains interesting even after the original author has accepted an answer and additional answers are always welcome.

Indeed, if we were to treat acceptance as shutting down a question, then we should not close duplicates.

  • @dmckee: if 17 months isn't late, what is? 24 months? – John Saunders Feb 27 '10 at 22:41
  • 22
    @John: Nothing is late. Because questions don't expire. There is no end date. I should say that at first that language didn't bother me, but it must have been niggling at my subconscious, and when I came back from my run and looked again at the title I just didn't like it. Certainly an answer months (or even days) on is unlikely to help the OP much, but so what? Questions are also there as a reference and delayed answers may help folks who come along later. – dmckee --- ex-moderator kitten Feb 27 '10 at 22:46
  • maybe "late" is the wrong term. Maybe "stale" is a better one. 17 months out, it's harder to know whether the answer actually answers the question. The OP is no longer involved. – John Saunders Feb 28 '10 at 1:32
  • 4
    I agree wholeheartedly. Ideally the question space should increase more slowly after time. That is, SO is not a support ticket site, but a question and answer database. Old questions, if relevant, should continue to be upvoted, bad answers downvoted, and hopefully edited to make them easier to find for "late" people wanting the answers to the same question. Perhaps if a "better" answer comes along later, the community should be able to re-accept that answer, and the "stale" questions convert to CW after X time. – user133393 Aug 3 '10 at 18:10

I don't see a problem with late answers on a question. In fact, quite the opposite -- assuming the answer is of "average" quality, more responses provides more information, which is a good thing, isn't it?

We should judge each answer on its own merits -- if it's bad, vote it down, regardless of whether it was the Fastest Gun in the West or arrived a year later.


I see nothing wrong in promoting one's own blog or product as long as it's relevant and it may be useful to the asker.

  • It's less likely to be useful to the asker if the answer comes 17 months after the question. – John Saunders Feb 27 '10 at 22:39
  • 10
    @John Saunders that doesn't matter really. SO's stated goal is to be a searchable repository of questions and answers. Thus, it's great if older questions get answered as well, no matter whether it still helps the original asker or not. – Pekka Feb 27 '10 at 22:50
  • Ok, so the developing consensus here is that it doesn't matter that there was an accepted answer 17 months ago? – John Saunders Feb 28 '10 at 1:31
  • I tend to flag references to commercial products as spam when posted by new users, unless the user discloses (s)he's actually working at the referenced company. True, the product might in fact be relevant, but I just dislike spam too much. (Often, on SU, these relate to PDF printers, Word-to-HTML converters and password recovery tools.) – Arjan Feb 28 '10 at 3:46
  • @Arjan: how is it spam if it's relevant? What difference does it make if the person is the owner of the product or not? – Thomas Bonini Feb 28 '10 at 11:08
  • The difference, to me, is how they advertise it. If they're trying to fool me in thinking they're actually users, then I flag. I know, I cannot be sure the user is not a real user... – Arjan Feb 28 '10 at 20:13
  • 4
    I'm with Arjan on this. There is a judgment call to be made. If the "answer" is not responsive to the question, flag as spam. If the answer consists only of a link, flag as spam. Otherwise, read the context and check the users other actions. Repeated posting of their link is a bad sign. Explaining their relationship with the product is a good one. Listing trade-offs between the suggested solution and other possibilities is also a good sign. – dmckee --- ex-moderator kitten Feb 28 '10 at 20:45
  • 1
    An example, probably gone by the time you read this: Do you need convert your PDF files to Word documents?If you do, here is a nice suggestion. [url] at superuser.com/questions/37726/convert-pdf-to-word-document/… and Download a free pdf to word converter named named Nemo pdf converter (which, however, is not free at all) at superuser.com/questions/104656/… both by an old user superuser.com/users/27276/nemo who even has the same URL in his/her profile, but no disclosure in the posts... – Arjan Mar 1 '10 at 7:52
  • 1
    as another example of self-promotion: it's on-topic, link-only, not linking to a product but a procedure/article at the user's blog: superuser.com/questions/114445/… ... this user should be encouraged to expand the answer to provide at least a summary of the blog post. it is self-promotion, but it's not spam. – quack quixote Mar 1 '10 at 9:59

If the late answers are spam, flag them. If they're good, upvote them, and help some guy get a Necromancer badge. I don't see the problem.


IMO, they should just be left alone. Since SO is intended to be a resource, more information at a later time is fine, especially (and I know this doesn't apply to your two examples) if the language changes and older answers no longer work.

It doesn't matter if this helps the OP. Let's say I have the same (or a similar) problem two years later. If we didn't allow the 17 month late answer, then I would be stuck with two year old answers that may or may not work. Even if they do work, the 17 month late answer may provide the solution that fixed my problem, since my problem is probably not identical to that of the OPs.

However, if they are spam (because of the content), then they should be flagged as such.


As long as the answer adds value, it doesn't matter when it's posted IMO.

Don't forget there is always the possibility of a question being washed ashore a year later by the community user. It doesn't have to be somebody who actively searched for unanswered questions. And even if - so what? Sorting out old questions should be encouraged, shouldn't it? I admit it may be fishy if a product is actively promoted, but the blog post I find totally innocent.


As the author of the first post (with a self-blog reference), I wanted to explain why I answered. Certainly, as someone new to Stack Overflow, I see that I didn't follow the normal naming conventions. My perspective as a Google user who hits Stack Overflow for answers is pretty different than an active Stack Overflow member. I'll absorb this feedback to answer differently in the future.

I think Stack Overflow is an awesome resource that has, in a short period of time, redefined how programming information is shared. Its tagging system has a side effect of greatly increasing its value to search engines. As such, the popularity of Stack Overflow questions can encourage the notion that a particular answer is the final word on a particular subject. It can drown out other voices on a particular subject.

In my case, I wanted to solve a particular problem that is very likely to result in a search engine hit for a question which has a good accepted answer. As a technical aside, the accepted answer was spot-on for self-hosted WCF services but not directly applicable to IIS hosted WCF services, which is what my scenario was.

  • @John: thanks for responding (that's why I pointed this question out to you). @Community: what can we do to address John's concerns? – John Saunders Feb 27 '10 at 22:40
  • @all: is this perhaps one of those cases where it's good to ask a new question and then to answer it yourself? The original question was finished, but @John: had more to add. Maybe a self-Q&A would be a good way to handle this? – John Saunders Feb 27 '10 at 22:42
  • See also "Are answers that just contain links elsewhere really “good answers”?" at meta.stackexchange.com/questions/8231/… – Arjan Mar 1 '10 at 7:46
  • 5
    I'm the person who wrote the original question. I frankly don't understand the hostility toward John. Although I had already accepted an answer, I rather enjoyed reading his blog post and I'm glad he submitted it for myself and others to see. Thank you, John. – Kilhoffer Mar 1 '10 at 13:18

You assumption seems to be that Stack Overflow is predominantly built for the Question asker. On the contrary, the asker is merely one person, and the acceptance of an answer shows just the opinion of one person.

Questions and answers are read by tens, hundreds, or sometimes thousands of people. So, if an answer adds value to the discussion, or even if the answerer just wanted to research a particular topic and write it up, or even if they just feel like answering, they should.

There's really no such thing as a late answer.

Spam is an independent issue. Early answers can be spam too. The timing of an answer doesn't have much, if anything, to do with its spaminess.


Answers that are only a link are always problematic and should be strongly discouraged.

But answers that contain a link and reasonable description/summary are (generally) perfectly acceptable, regardless of how early or late they are added.


How do you earn a Necromancer badge without adding a late answer?

  • John. A late answer yes, but not to a question with an accepted answer. That's what I meant by late - way after the accepted answer. – John Saunders Feb 28 '10 at 23:49
  • 4
    Sometimes accepted answers are downright wrong, and they're generally fixed in time. The best answer to a question in March '09, one well worth being accepted, may not be the best answer in March '10. – David Thornley Mar 1 '10 at 15:16

Two words: stale links. I have found any number of answers which were extremely helpful at the time the OP was involved, but later were ... to continue the terminology ... stale because the main component of the question was a link to an external site or product. Late answers can be very useful, especially as newer libraries / products are released.


As you seem to feel this way because an answer was already accepted: accepted answers can be changed at any time. (Unless there was a bounty.)

  • I felt that a 17 month old accepted answer indicated that the original poster was well satisfied with the answer. In that case, I questioned who the new answer benefited. The OP was no longer able to tell us whether the new answer was helpful, so who did it help? – John Saunders Mar 1 '10 at 8:30
  • 1
    (@John, like people who cannot have their duplicate asked, but that has been covered in another answer.) – Arjan Mar 1 '10 at 8:53
  • I'm sorry, I don't understand what you're saying. – John Saunders Mar 1 '10 at 10:25
  • @John, I mean it might have helped (or: might one day help) people who have a similar question (but might like a different answer), but cannot ask that question without it being closed as a duplicate. – Arjan Mar 1 '10 at 11:29

You must log in to answer this question.

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