I think that about Subversion, which is really just a versioned filesystem; the git command line is pretty complicated though.

I don't know if the versioned filesystem metaphor fits Subversion that well -- clients do not store history after all. It is usually what Git evangelists praise about their software; you can use Git as a simple versioned file system if you wish (git add/commit/log and you're all set).

Subversion is literally a filesystem; that is all it is. Most of Subversion's problems come from the fact that the metaphor doesn't fit the task very well.

The fact that Git clients store full histories is one of the things that makes it less like a filesystem; it is, instead, a system for managing overlapping patch streams.

Check out the core svn C APIs (nb: our product is built on them); it's basically an open/close/read/write/readdir system with one additional "history" option (which history snapshots an entire filesystem, not just individual files (like CVS) or patches (like git).

in many cases with a large project it's faster to work with a GUI. You can navigate complex hierarchies more quickly.