Often(basically always but I'm not 100% sure) when I download papers from any journal they have random letters and numbers as their title, this is quite annoying and I have to rename the document to have a descriptive name. Why do this files have these random filenames instead of the titles of the papers?

  • 5
    There is no requirement that published paper titles be unique. There is a requirement that the filenames be unique.
    – Jon Custer
    Sep 13, 2018 at 18:05
  • I mean the title was just a suggestion, adding the authors in the end might solve it or something else like adding a numbering.
    – Hakaishin
    Sep 13, 2018 at 18:55
  • 1
    When combining with bibliography management software this need perhaps no longer exists. Also, individual's naming preferences are not at all the same, maybe it's easier just to keep the random chain of characters. Yet, if you're interested to rename them, so programming at the user end may like this may work? Sep 13, 2018 at 19:12

1 Answer 1


The filename is some sort of encoding that is compatible with the database in which they keep the papers, it might have a real world meaning or not. It is optimized for their own processing. I seriously doubt that they use a flat file system to store the documents.

But in general, trying to use a paper's title as a file name won't work anyway, since some characters that might occur in a title aren't accepted in filenames on every operating system. Colon and slash, for example, are problematic. There is also the problem that a paper might need to go into the database during the acceptance/review process in which the title of the paper isn't finalized.

In addition, while we like the titles of papers to be unique, that isn't always the case. So, some disambiguation needs to be done.

What you have, actually, is a reference to the paper that represents its "identity". This is similar to a pointer or reference in a programming language. The thing referenced can change arbitrarily, but its identity doesn't. This help guarantee that they don't lose it.

  • 1
    Last point besides, this sounds like it is practical reasons why it is not done, but many other platforms and services allow for smart filenames. This answer makes it sound like all digital files could only have bizarre names, which is obviously not true.
    – Hakaishin
    Sep 13, 2018 at 16:18
  • 3
    @Hakaishin but when I put a file on my system, I know pretty much what works and what won't. I don't know what will work on yours, nor certainly everyone's. True that systems are getting smarter, of course.
    – Buffy
    Sep 13, 2018 at 16:22
  • @Hakaishin Yeah, but many platforms don't allow for that. There are probably academics out there running 10 year old OSes. Sep 13, 2018 at 21:24

Not the answer you're looking for? Browse other questions tagged .