I have noticed several posts with links to pdf files where IP address is used instead of domain name the url.1 (I will list them below.) Several of them seem to be dead. Moreover, this type of url seemed unusual to me. So I wanted to ask a few thinks about them:
- Do they work for other users? (When I tried the links, I did not get the files. But it's possible that I am the only person with the problem.)
- Is it possible to say whether they worked at the time when they were posted? Is it possible that the links work (or worked) only for some users?
- Are there some advantages to posted such links rather then using the more "usual" form? Is there some plausible explanation why somebody might have posted such links by mistake?
Here are the links that I found:
- There are several (thirteen) posts containing url's starting with 129.81.170.14 [All fixed 21 Aug]. For example, http://129.81.170.14/~vhm/papers_html/final-bell.pdf, http://129.81.170.14/~tamdeberhan/gregweb.pdf, http://129.81.170.14/~tipler/theoryofeverything.pdf. Neither of those links works for me, but the links http://dauns.math.tulane.edu/~vhm/papers_html/final-bell.pdf, http://dauns.math.tulane.edu/~tamdeberhan/gregweb.pdf, http://dauns.math.tulane.edu/~tipler/theoryofeverything.pdf work fine.
- I found an answer which links to http://130.44.194.100/proc/2000-128-05/S0002-9939-99-05670-1/S0002-9939-99-05670-1.pdf. This links does not work for me, while https://www.ams.org/journals/proc/2000-128-05/S0002-9939-99-05670-1/S0002-9939-99-05670-1.pdf works. [Fixed 16 Aug]
- There were 3 posts with urls containing 131.220.77.52. (One of them was recently fixed: https://mathoverflow.net/posts/215306/revisions.) For example, the link http://131.220.77.52/lueck/data/ictp.pdf. Unlike that one, the link to https://www.him.uni-bonn.de/lueck/data/ictp.pdf works for me. [All fixed 21 Aug]
- In one answer I saw a link to http://193.224.79.10/~p_erdos/1960-09.pdf. A working link is http://bsmath.hu/~p_erdos/1960-09.pdf. [Fixed 16 Aug]
- I found a question linking to http://146.163.152.131/research/papers/sdrbounds.pdf. The link http://www.siue.edu/~aweyhau/research/papers/sdrbounds.pdf seems to work fine. [Fixed 16 Aug]
- There is an answer linking to http://194.42.1.1/~nickp/lectncm.pdf. The corresponding url could be http://www.mas.ucy.ac.cy/~nickp/lectncm.pdf. [Fixed 16 Aug]
- There are also a few "IP-type" links in some post which do work (at least for me): http://129.187.111.185/~dieter/football.pdf, http://129.187.111.185/~dieter/duke2.ps, http://129.69.211.95/pdf/mit/lcs/tm/MIT-LCS-TM-560.pdf, http://163.14.246.20/mp/pdf/S13N26.pdf.
1I was playing around with checking MO posts for dead links. This was based on YCor's suggestion in a comment to another question. Originally I only checked jpg files, at that time I posted some details about what I triead in chat. Recently, there was also a discussion about archiving externals links from posts on SE sites on Meta Stack Exchange: Automatically have links archived in the Wayback Machine.
129.81.*.*
is Tulane University.130.44.*.*
is AMS.130.220.*.*
is University of South Australia.131.220.*.*
is Uni-Bonn.129.69.*.*
is Uni-Stuttgart.129.187.*.*
is Leibniz-Rechenzentrum (Dagstuhl?),146.163.*.*
is Southern Illinois University Edwardsville. $\endgroup$129.81.170.14
and two with131.220.77.52
. I can slowly edit these over the next week. $\endgroup$129.81.170.14
. I'm taking the liberty of improving the references where possible, since opaque see this link references force users to click through merely to see if they already know what is being linked to. $\endgroup$