November 27, 2007
If You Click On A Broken Link, It's Not Your Fault...
Who's responsible for fixing broken links on an intranet website? Not the visitor, that's for sure - but how often are you made to feel responsible, as if the broken link was in some way your fault..?
A couple of months ago, I clicked on a link from one OU intranet page to a page on another OU subdomain (for example, it went from "intranet.open.ac.uk/thisPage.html" to "iet.open.ac.uk/thatPage.cfm").
The link was broken, so good citizen that I am I reported it to the owner of the second page telling them where I had found the broken link in the first place.
My thinking was this: if the page existed once and was moved, a permanant redirect could be put in place; as we all work for the same institution (?!), the owner of the second page can tell the owner of the first page where the link should point to; the error page was horrible and made it look like maybe the error was my fault - maybe it could be replaced with something friendlier? And finally, the error page told me to report the error, although it didn't provide an email link or a form to make that reporting easy...
The response was as expected - "the link's not ours, so it's not our fault. I'll pass it on to the page owner, harrumph".
I just tried the original intranet page again, and the link on it is still broken (it's the link to the CREET website on this intranet page; hovering over the title of the Centre looks like it's a link, but it isn't; the link is below the centre description ;-). The error page is a little friendlier now, and if you're into such things, you'd notice it as a customised 404 page ;-) It also includes a link to report the error (so I did.... again...)
Now I know that there are all sorts of issues about what to do with error reports like this, but I can't believe there isn't a better way to try and repair broken links on an intranet - or inter-intranet (you know, a set of internal, inerconnected networks within the same organisation).
When the server receives a request for a page, and it can't find the page, it generates an internal error that can be used to present the user with a particular error page, which may be humourous, or may be pretty bland.
It can also tell whether the request came from a user clicking on link on another web page, as well as the location of that page. So it could in principle generate a report for the owner of that referring site telling them that one of their pages contains a broken link, and would they like to chat about it...
Or of course, you could just let the links continue to rot...
Posted by ajh59 at November 27, 2007 03:42 PMI don't think I quite understand what you're suggesting. The 404 handler sniffs the referrer, and if it's from the same domain, it prepares a report for the owner of the referring site, fine ... but how does that report get to the owner of the referring site? There's no way of a script knowing who that is. (Ghods know it's hard enough for a human to find it out sometimes.)
Posted by: Doug Clow at November 27, 2007 04:26 PM"The 404 handler sniffs the referrer, and if it's from the same domain, it prepares a report for the owner of the referring site, fine ... but how does that report get to the owner of the referring site?"
If you get a referrer report for a broken link to mysite.com/thisPage.html from linksrothere.com/whoCaresWhereTheLinksGo.html
you could chance an email to admin@, info@ or postmaster@linksrothere.com
However, internally, it *shouldn't* be that hard to find out who a site admin is (though finding the page owner might be...;-). Especially for a group that claims to run an award winning knowledge network...;-)
I suppose it boils down to how much you care about whether or not people are linking in to your site, and how you handle broken links that do come in.
At an institutional level, I guess someone has a stake in the amount of time that gets wasted following broken internal links? And the subtext of the post was - WHO is responsible for fixing broken links on interoperable internal networks?
"There's no way of a script knowing who that is. (Ghods know it's hard enough for a human to find it out sometimes.)"
Yeah, you're right of course - let the links rot... No point trying to work out a way of getting 'em fixed... ;-)
In the case referred to in the post, my assumption was that the second, linked to site was to blame because I assumed they had moved the original link (rather than the link being broken from day one)... If you can keep your URLs persistent, then find a way of only publishing persistent URIs ;-)
[I realise of course I am setting myself up for a flurry of emails about broken links in OUseful.info posts ;-) And the thought of what to do if I have to move the OUseful.info blog to another server/another domain keeps me awake at night!]
Posted by: Tony at November 27, 2007 05:32 PMWhen web 3.0 arrives, we will have smart self-correcting links which never rot.
Oh wait, that's called Google, which is how I tend to find things on our intranet ;-)
I'm not sure I'd necessarily class that as a 'broken link' as from your screenshot it's an error in the server-side code (in ColdFusion in this case) rather than an HTTP error. So the page is actually meant to be there, even if it's not working. Whether it'd be worth removing the link would really depend on how soon the owner of the broken page is planning to fix the bug I guess. It's often not easy to find out who 'owns' a particular page though of course.
Posted by: Juliette at November 27, 2007 09:32 PM