This is a fair point. Here's my thinking: Just because a paper is hard to find does not mean that it doesn't have valuable data or ideas. To give an example, I

This is a fair point. Here's my thinking: Just because a paper is hard to find does not mean that it doesn't have valuable data or ideas. To give an example, I spent a lot of time tracking down Russian papers in my field because I found their theory to be better developed in some areas than the west's. Should I not cite these papers because most people won't be able to find them? I don't think so. What they say is unique. There is no alternative. Instead, I translated the most important papers into English. In a conference paper I'm writing right now I cite quite a few of these papers and I've tried to be extremely specific in their citations, going so far as to include OCLC numbers that I used for interlibrary loan. (Librarians will know what to do with this information.) What's easily found today may not be so easily found in the future. Today, I find that many researchers can't find a paper that's not available online but is available in any decent academic library. It's also possible that a hard to find paper is cited wrong or cited ambiguously, and having the correct, unambiguous citation is all one needs to find the paper. To address papers which are practically impossible to track down, as gms7777 suggested, at least people can contact the author of the paper citing the hard to find paper in this case. (For the Hinze paper, as I recall Shell had a NDA as a condition of downloading the paper from their system, so I can't share that one.)