I really like that one, If we weren't the April 1st, I may have considered this a real project.

A first-class c++ troll, I admire the writing.

This seems overly complex to me. I would have preferred xmlsh.

Ah, but should the command flags be sub-elements or attributes?

Both, obviously.

What I find especially annoying about these April 1st articles everywhere, is the people who think that they're so uniquely clever that they've figured out that it's not real, that they need to post a comment saying 'April 1st!' Duh, no shit Sherlock?

(I realize the irony in that this post can very reasonably be interpreted as me announcing that I've figured out that this one isn't real...)

I'm waiting for whish: a bash-like shell with Whitespace syntax.

There is already the C++ shell for root: http://root.cern.ch/drupal/content/cint

True, though it's targeted at the sightly different angle of being a REPL for an analysis environment, rather than being a command-line shell for general CLI computing. So it's more of a competitor to something like R, or to the way Python is used as the top-level of SciPy.

Yes, maybe most sadly, the CERN guys aren't joking around... :/

Disturbingly reminiscent of powershell

In what way?