Just to give the context I had the same doubt of the our coleague @JOB in this Thread: "Unable to Access PDS , ok its was solved. I have one question which is related to: Why when us are on that envirnoment Linux, from the command TSO OMVS, we aren't able to see us Dataset's, PDS's? Or Is that possible?
You have to understand that z/OS-datasets and OMVS files live in two different worlds:
z/OS datasets do have a name that consists of a series of qualifiers but are not really organized in a hierarchical manner, they are distributed over a cluster of (virtual) disks and you have a couple of catalogs to find them again. There are no such things as directories: while you might have a dataset named MY.COOL.DSN, there might not be an object that is called MY.COOL or it might be an ordinary dataset as well.
OMVS files on the other hand live in a filesystem that has a hierarchical structure. Each file might reside in a directory, that might be in another directory etc. In total you end up with a tree of directories with a single root-directory and files as leaves.
To realize this in OMVS you create z/OS datasets that contain hierarchical filesystems - either HFS or z/FS - each with its own directory-tree. Upon installation one of those is mounted as the root-filesystem and addressed via / and you might mount other HFS or z/FS filesystems on any point further down the directory-tree, so that it is added as a subtree.
If you are familiar with Linux you can compare the HFS and z/FS datasets as disk-partitions that can be mounted in your system.
Long story short: when navigating via cd and ls you are moving through the directory-tree that consists of all the mounted z/FS and HFS datasets, but there is no defined place that contains the ordinary z/OS datasets - and there can't be since they are not organized in a tree-structure.