And then you have to put a filesystem on it, which has its own metadata – file attributes and folder/file names and so on. If you use NTFS you lose at least 12.5% to the metadata so now you’re down to 11.8 GiB. 😛
As an amusing side note, I once came across a joke compression program that could compress any data down to zero bytes. It did this by creating directories filled with zero-sized files whose filenames contained the actual data of the file in question.
If you right-clicked on the folder and asked the OS how big it was, it’d report 0 bytes. But of course all that data still had to be stored somewhere, in the metadata of the filesystem.
That’s part of why I use du on Linux instead of df/ls -l to figure out file/directory/partition usage. The former figures out actual size on disk, whereas the latter ignores metadata like the list of files in the directory.
And then you have to put a filesystem on it, which has its own metadata – file attributes and folder/file names and so on. If you use NTFS you lose at least 12.5% to the metadata so now you’re down to 11.8 GiB. 😛
As an amusing side note, I once came across a joke compression program that could compress any data down to zero bytes. It did this by creating directories filled with zero-sized files whose filenames contained the actual data of the file in question.
If you right-clicked on the folder and asked the OS how big it was, it’d report 0 bytes. But of course all that data still had to be stored somewhere, in the metadata of the filesystem.
That’s part of why I use
duon Linux instead ofdf/ls -lto figure out file/directory/partition usage. The former figures out actual size on disk, whereas the latter ignores metadata like the list of files in the directory.