Imagine if filesystems didn't have path hierarchies. Every directory would be represented by an integer in a flat hierarchy.
Whenever you want a new directory for your application feature, you'd pull request on someone's Github repository to reserve a new integer so that others don't use the same for a different purpose.
If you don't know the meaning of a directory number on your system, you can always check that one Github repo.
There's also no filenames, so you can't replace an existing file, only add to the same directory.
It's wonderful that paths exist.
Imagine if filesystems didn't have path hierarchies. Every directory would be represented by an integer in a flat hierarchy.
Whenever you want a new directory for your application feature, you'd pull request on someone's Github repository to reserve a new integer so that others don't use the same for a different purpose.
If you don't know the meaning of a directory number on your system, you can always check that one Github repo.
There's also no filenames, so you can't replace an existing file, only add to the same directory.
It's wonderful that paths exist.
Login to reply
Replies (5)
That is why the semantic web uses URIs as identifiers.
The horror… I hope this doesn’t exist in practice 😰 wait… 🧐
Reminds of the way PalmOS and BeOS used to work… “files” were essentially content databases. So,e neat stuff was starting to happen with the concept of “app” around the time PalmOS faded. Could have been some interesting address-space-meets-transformative-content-bits
Isn't that a bit how INODEs work?
I think lot of filesystems look like that under the hood.
Would you consider git a "pointer-based" filesystem with no hierarchy?