103

Stack Exchange runs a bot that validates all the external links in our network of sites. This bot performs HEAD requests in a heavily throttled way. This bot was created to combat link rot in our network. Links are tested once every 3 months. If you notice any issues with this bot please email team@stackexchange.com The proposed operational mechanism (...


33

The Wayback Machine now has an availability API http://archive.org/about/wayback_api.php It might make sense to incorporate this into the crawl. If we can't find the original source link in the crawl, ask Wayback if they have a copy of said link, and if they do we can fix the post ourselves without bugging users about it. The crawl is of course still ...


31

humbly accepts compliment I suspect it's caused by users who copy/paste a link from elsewhere, but accidentally select too much text, more specifically an extra space in front of the link. Like this: Do you see the difference? I don't, and that's exactly why these broken links happen. The insert dialog above produces the following link: enter link ...


28

If the link is broken, it is helping no one. It is best to replace it with a canonical one from elsewhere, but if that doesn't exist, an archive.org link is absolutely fine. Your apprehensions are a bit misplaced, IMHO. You can't guess as to why a page is no longer there and people who do post things on the web should know that "the web doesn't forget".


26

If the answer contained so little information without the link, it wasn't a good answer to begin with and probably deserved a downvote at the time. Now, since the link is broken (unless you can fix it, in which case please do!) it should be downvoted as it's now even more useless. The downvote tooltip says "this answer is not useful". It doesn't say "this ...


24

Good eye that this could improve linking, but I downvoted this - SE should not get involved in cat-and-mouse games with arbitrary sites to defeat their policies. If Quora really is a significant portion of our outgoing traffic, then we are a significant portion of their incoming traffic, and they'll do the "cat" part. Secondly, hacking competitor sites ...


23

We should complain to Oracle that they messed up their redirects. Currently, http://blogs.sun.com/chengfang/entry/why_am_i_using_netbeans redirects to https://blogs.sun.com/chengfang/entry/why_am_i_using_netbeans, which redirects to https://blogs.oracle.com/roller-ui/errors/404.jsp (which despite the filename, responds with HTTP 200). (checked using wget '...


23

It's possbile that neither the user who cast the "invalid flag" flag nor the moderator (maybe precisely because of the "invalid flag" flag) were aware that the link is defunct. The flag queue is huge and there are only a few moderators. In cases like this, rather than just casting a NAA flag, you should flag as Other and explain that the link is now dead ...


14

This is done. Big thanks to Jarrod Dixon for making a tool for doing simple URL search-and-replace across multiple posts!


14

You cannot fix it yourself, however, a moderator can. Flag the post for moderator attention and explain the situation ("link in text xxx is broken, could you replace it with yyy or remove it?"). Now, if it's historically locked, as this question is, you almost did the proper thing. Ask on the Meta of the corresponding site to bring attention to the question....


13

If the answer is link only and hence now worthless you can flag it and put a comment in the "other box" to make it very clear that this is now worthless. If the answer is free-standing and the link is only a small part you can edit the post to remove the broken link or correct it, nothing is lost and there's no need for any moderator involvement at all. ...


13

This depends entirely on the audience of the site. On sites like Biology or Chemistry, a paywalled link is not an issue for a large portion of the users. Anyone working in the field needs access to the scientific literature anyway, and generally has access through a university or company. In a professional scientific setting there is also nothing like a ...


12

The broken links have been fixed across the network. Thanks to everyone who pitched in.


11

If at all possible, we should try to improve answers, so I tried to use the Wayback Machine to see if I could at least get a copy of the old content and at least put this "answer" back into it's previous state of being helpful. However, a robots.txt file on that server prevented the Internet Archive from parsing it while the link was still active. Thus, in ...


11

So, there where two issues here. 1) Someone (who I will not name for the sake of all involved) ... may have not chkconfig on'd the named service when he/she/it rebuilt the NY DNS servers to upgrade them to CentOS 6.x This caused two of our 4 DNS servers to not respond to any requests when they where rebooted for patches yesterday night. We should have ...


11

The W3C Link Checker test executed on the Leagues page gives server side errors in the following seven links, which I have manually verified to be broken: http://stackexchange.com/leagues/185/week/martialarts http://stackexchange.com/leagues/183/week/smugmug http://stackexchange.com/leagues/68/week/writers http://stackexchange.com/leagues/215/week/patents ...


10

Starting with the next build, we'll be relaxing this throttle for established users. Broadly speaking, the higher your reputation the more edits you'll be able to make before running into this anti-vandalization cutoff. Past a certain point, the limit will be removed altogether. A flag may still be raised to call someone in to look at an unusually long ...


10

As of 2013-10-02, the redirects for issues and pull requests in the old Bootstrap repository appear to be functioning, so no edits are necessary. Generally, there's supposed to be a redirect from the old GitHub repository to the new one in cases like this. This wasn't working after Bootstrap's transition to its new organisation, so someone opened a relevant ...


9

Good idea, working links are important. Fixing links makes the StackOverflow experience better. Not sure if this particular case is worth the effort, but 800 links sounds like a lot. Shame there's not an API that allows you to script this :-)


9

Personally I am not sure whether this is having the expected result. I am not sure about that company / site, but it is possible that Google, Microsoft, Apple and others are more reliable in keeping up than it is with that site. In fact, you are swapping the dependency to those sites (which will be 90% of the total urls in SO for example), to an external ...


8

I just fixed it by adding an explicit anchor to the Wikipedia article, so that the old URL will lead to the correct section again. I also fixed the redirect from Meek's method to Counting Single Transferable Votes so that just linking to http://en.wikipedia.org/wiki/Meek%27s_method without any section anchor will lead to the correct section. In fact, this ...


8

These were excluded during a recent build system change - they will be back shortly.


8

Like Bart said in a comment, this shouldn't be too much of a problem. If a link dies, the content should already be included in the answer, so that the link should not be necessary anymore. If you find that there is a widespread problem of "link-only" answers on your SE sites, you should either edit those answers to include the important information, or ...


7

The link isnt broken; StackExchange is just using the feature a little bit differently then normal. They provide the content at the top of the page and in the normal place instead of just jumping to it on the page. Note the link you provided shows the requested content in a highlighted box at the very top of the page and also more near the bottom which is ...


7

You might have noticed (since it is making edits all over the network) that I wrote a bot which performs this task on old posts where the link has broken (because of not existing anymore, or not available via HTTPS, which is a requirement nowadays). It will add a link to the original URL, either in the form of a "source: example.com" blurb under the post, or ...


6

The redirect is back up and working, give it a try: http://odata.stackexchange.com SEDE was totally rebuilt with our Oregon web tier, and the old binding for odata.stackexchange.com simply didn't survive the move...it's just been corrected.


6

Looks like someone simply forgot to add that link. It's pretty obvious what that was meant to link to, so I fixed it here; all privilege pages are wikis on MSO. The change will propagate to all other Stack Exchange sites in… actually I don't know, this is done manually.


6

My typical thought process is: If the answer provides some context, but not much, and most of it is in a relatively obscure link: leave a comment about the broken link, possibly downvote depending on if the answer is otherwise well-formatted/good/etc. If the answer contains a link that is easily findable (or known to have moved due to version updates): edit ...


6

Ok. I just copied most of these over from MSE verbatim. Not the most elegant solution, but... No more broken links. Of course, most of the tags will be deleted in 24 hours and I'll have to re-create them after we migrate questions that actually use them. But, the concept is sound.


6

Jon recently moved his blog, and there are some links around that have not been moved. I thought we did this one - we have now.


Only top voted, non community-wiki answers of a minimum length are eligible