Hacker News new | past | comments | ask | show | jobs | submit
> I think there is always a question of where to file something, like a book about the history of gardening in England could be filed under gardening or the history of England.

i understand the pain point, but:

1. modern libraries usually have an information system that precisely tell you a) if they have the book you're looking for at all and b) where it is in that specific library

(the two city libraries i go to have a computer for that, or you can ask the librarians)

2. every time i search for a specific book i serendipitously run into other interesting books, and that's nice (so i've got that going for me).

side note: i recently was looking for "simulacra and simulation" by baudrillard and asked chatgpt under which section to look for it. scanned for a while, did not find it, noticed some interesting other titles, then finally went and checked the information system and that specific library did not have it at all. so lesson learned is: don't bother much with the classification, just use the digital information system

The problem is that it completely discards the other, also important, use of a library: discovering books that you did not know about which are slightly related to what you went looking for.

If you go look for Metamagical Themas by Hofstadter, you already know to search for Hofstadter, but maybe you hadn't heard of Martin Gardner or Raymond Smullyan or Persi Diaconis, all of which you could conceivably find on the same shelf.

This sort of serendipity is lost in catalog systems, even keyword systems -- but it is one of the best features of Wikipedia, and one which the Wikimedia Foundation seems determined to stamp out.

> This sort of serendipity is lost in catalog systems, even keyword systems

I kinda see your point and that’s why i think information systems should be deployed alongside catalogue systems.

Also, what about recommendation services?