It's a pretty good idea, but it will work only for low/normal usage.

Every file-system has limitations regarding how many files it can handle and also how many folders a folder can have, so if each folder is used to store a unique ID... pretty soon the file-system will be collapsed with them, well.. actually not pretty soon in most cases.

The limit depends on the OS, usually is between 20K - 65K sub directories, therefore that is the limit for upload id's in this case.

I have think on a way to store files on a most efficient way, it consists on directory population. The upload dir is separated by categories (like current year, or index number, etc) and then, this folder contains child categories and then the images, each folder populated with the recommended amount of files (1000, 5000, depends on the file-system). This will suit perfect for large storage because we can actually store the files in a not abusive way.

But most than a script for large storage, Chevereto trends to be used in a personal way... Which means low usage, not so neccesary to have this kind of limits in mind, etc. So, a personal mode can have this kind of storage method as an option for those that want to preserve the original file names.

Yup, a personal mode have been always in mind.

Actually, the whole filemanager thing and a private mode are just the first steps into this direction.

As I said before, that function will be implemented on 3.0

2.1 is under development and it seems that from 2.1 we will jump straight forward to 3.0, but 3.0 develop will take pretty long to start because after 2.1 I will move the whole thing to IPB and I will create a Chevereto fork called Quickty (general file upload).