Information Research Weblog

You may recall that we had a little discussion on IR-DISCUSS a while ago on the problem of dead links in electronic journals and on the probable impossibility of maintaining such links. After all, print journals that cite dead links don't issue addenda, do they?

However, I did a check, using Xenu - an excellent program, by the way - and found that there are now 2,728 links on the journal site, both internal to the journal and external from papers to the Web. 77.79% of these are still valid, which is pretty good, I reckon. 9.16% of the rest are of a type that Xenu skips (e.g., mailto links); 9.27% were not found and 0.48% were 'forbidden request'. 1.65% reported 'no such host', 1.21% were timed out and the remaining reasons covered a total of 12 links.

I intend to get rid of the internal bad links and we'll see how many that leaves that are dead externals.

If anyone has a good idea about automatic correction of links, I'll be pleased to hear about it. However, since it will have to cost no money and need no technical skills, I suspect I'm looking for the impossible!