Since one of the main goals of SO is to provide a comprehensive Q & A repository (and not just a question and answer forum), how should we mitigate the degradation of Stack Overflow as external content dissapears or changes such that the links no longer point to answers of the question they were addressing?

I have two ideas, but I am hoping others have ideas as well, and I am curious to see if this is being addressed.

My ideas:

  1. Allow external links only for backup to answers, not the actual answers themselves (somewhat hard to police)
  2. Automated process whereby StackOverflow saves off a copy of the external content, and changes the link in the post to point to StackOverflow's version
  3. Automated process that alerts "volunteers" to broken links, who can then go manually fix them or delete them or whatever

Part of the problem of degradation is that it is two-fold.

First, degradation may occur when links break due to content dissapearing, in which case the link just doesn't work anymore.

Second, degradation may occur if the content on the linked page changes to the point where it no longer answers the question.

Ideas, anyone?

  • I've frequently thought that #3 was a good idea.
    – Gabe
    Mar 18 '11 at 3:15
  • 1
    Comments and down-votes on link only answers will encourage people to post the information on the site.
    – ChrisF Mod
    Nov 9 '11 at 14:35
  • I thought we already had #1 as a policy. If this isn't obvious, we've failed at the (admittedly difficult) task of policing the rule. Nov 9 '11 at 16:01
  • 1
    I think all has been covered in Can we have some tools to handle link rot?
    – Arjan
    Dec 12 '11 at 17:32
  • Not directly related to the Q, but a nice tool for retrieving dead links: Internet Archive Wayback Machine.
    – brasofilo
    Jul 30 '12 at 9:28

I was just about to ask about answers with dead links (similar to this one) when I came across this question. The most useful solutions in my opinion would be:

'Really Bad' link list

Implement in the StackExchange software a trawl of properly formatted links on all questions, answers and comments at sensible time periods specified by some sane rules (derived from the total quantity of links in a particular StackExchange site) and form a list of links (categorised by HTTP error codes?) which are considered 'dead'.

Appropriately privileged users can then view this list and visit the question/comment/answer containing the 'dead' link and (propose an) edit (to) the link which resolves it, or if unable to do so, alerts the user with an @ directed comment.

'Bad' links

Provide a specific 'flag' option which is for links which are technically not dead, but are unsuitable for various reasons (e.g. the server responds, but with something no longer relevant). The owner of the post containing the link could be notified and the link could be added to the 'Really Bad' link list.

These are just two ideas!


I'll address each of your points and add another:

  1. Yes policing will be hard. I don't think it is bad to have links to the answer even if the linked content contains the answer. While an answer should be self-contained, a link to an existing, more complete, or canonical source can be helpful, but I agree dropping a link as an answer and "running" doesn't seem useful.
  2. I think this is a great idea if it doesn't violate any TOS. The linked content certainly won't have the same licensing model. All that said, search engine caches and the Wayback Machine copy content. I don't know the full reasons, but I assume it amounts to fair-use. Perhaps if any of these are exposed as services, SE sites could make requests for the URLs to the cached content rather than store them. Even so, a temporary copy might be a good idea too.
  3. Absolutely a great idea.
  4. Change the color of dead links so readers will notice them, and possibly fix them.

You must log in to answer this question.

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