Hierarchical tagging is a good idea for a filesystem. The only problem with it is "where do you store that metadata?" In the individual files (requiring forma

Hierarchical tagging is a good idea for a filesystem. The only problem with it is "where do you store that metadata?" In the individual files (requiring format-specific support)? In the filesystem (requiring applications that know how to read and write it)? When KDE 4 was released it included Nepomuk (https://userbase.kde.org/Nepomuk https://userbase.kde.org/Nepomuk) as a universal tagging-and-search mechanism to do precisely this kind of thing. Unfortunately, not very many people use it and even those that do only use it in a very limited fashion. Partly because it only works in KDE apps and also because tagging all your files after-the-fact is a very time consuming process. Then there's this problem: What if I never had any trouble finding my documents? What will I gain by tagging everything? Tagging makes a lot of sense for things like photos and videos where search-by-content just doesn't work. For everything else you can just index your files and search normally. Another problem with tagging is that--if you're going to be disciplined about it--you're not really getting much benefit over just "staying organized" with directories/folders. What's the this-helps-me difference between a directory structure like, "Company X/Clients" or a hierarchical tagging structure such as "#company/#clients"? Sure, with tags you're not limited to the parent->child concept but ultimately--if you don't stay organized with a hierarchy--you'll end up with a huge mess that can only be navigated with an intelligent tag-based search tool.