/Wheel of Time
/TV Series/Wheel of Time
/Video Game/Wheel of Time
/Wheel of Time
/Wheel of Time (TV Series)
/Wheel of Time (Video Game)/Francis Bacon
/Francis Bacon (artist)
without having to answer anguish-inducing questions that would be raised by
/Francis Bacon
/artist/Francis Bacon
e.g. questions such as:
1. "Maybe we should put Lord Bacon under a category also, maybe `philosopher/Francis Bacon`"
2. "But he's also a statesman ... what about `statesman/Francis Bacon`? Is he more of a philosopher or a statesman"?
3. What about ordinary people (who happen to be involved in historical events) that have wikipedia entries, like George Floyd? Should he be assigned something like `person/George Floyd`? If so, should the two Francis Bacons be assigned `person/philosopher/Francis Bacon` and `person/artist/Francis Bacon` instead?
And so on
It’s a fundamental issue:
Hierarchical naming and categorization sucks.
In 95% of the cases, you want sets or even graphs rather than trees.
You _might_ be getting away with trees, because you didn’t encounter a case where the hierarchy breaks. But you’re being fooled by happenstance.
In the 5% of cases where trees are fine, you could also use sets or graphs.
This issue comes up all the time with all sorts of things, including urls and file structure.
The trick is to have a general “/type/name” format and to be as liberal and general as possible with what “type” means.
In programming, it’s useful to have namespaces for overall organization and conflict resolution. But it’s a trade off: you are nudged into a hierarchical ontology, with all the implied issues.
Wikipedia titles are free text (more or less). So they can afford not to introduce hierarchical naming and still have nice, easily addressable names without conflicts.
This avoids all sorts of problems. Most things simply can’t be categorized in a strict, hierarchical manner.
I realise this is somewhat tangential to your point, but shows how easy it is for innocent looking choices to end up creating annoyances.