#workflow
File systems on computers typically consist of folders, and each file must belong to one and only one folder. This is not how the ideas of a human being naturally behave.
An idea might be related to many topics. The topics themselves fit nicely into a tree-like structure: a subject has sub-fields, e.g. quantum mechanics is a sub-topic of physics.
It is therefore sensible to arrange ideas (and more generally speaking, any information) flatly inside one big folder, and upon creating each piece of information, attach it to various topics by tagging it.
Each object in SilverBullet has a main tag which is its “data type”, e.g. page, task, item, etc. Then, it has some extra tags (a Lua array) that indicate its topics.
For example, second_quantization belongs to topics #physics/quantum-field-theory and #physics/condensed-matter. So its tag is "page" while its tags is a Lua array of two items.
Tags are very helpful when you want to retrieve information from a collection of notes. Instead of searching or clicking through a folder, you can query this collection, just like asking questions to a human. For example, you can query for pages that simultaneously fall under two different topics. SilverBullet has Library/Std/Infrastructure/Query Templates that help you write queries quickly.
At https://silverbullet.md/Tag there is more information about how tags work in SilverBullet. The page custom_functions contains some Lua scripts which help query pages via their tags.
In the database version of Logseq, there are also two kinds of tags: the main tag indicating the type of “stuff” an element is, and some properties serving as extra metadata.