I made the suggestion to Mahmood to help develop a new GUI with tighter integration with a host; especially the docking part. I was looking for a minimalistic GUI to complement the more feature-rich QML equivalent, for times when you aren't looking for flexibility or features, but rather quick results.
The GUI, at least to my mind, could be a mere progress bar to the side of a host, like Maya, where a user could publish or validate things quickly.
Secondary, it would also enable those who are looking for a more lightweight alternative to Pyblish than QML. I've had users come to me and say they simply don't like QML or that they don't like the fact that the GUI doesn't stay on-top of their host. All reasonable concerns that a separate GUI could address.
And thirdly, I suppose, is to enable a separate flow of ideas that introduce a form of competition. Seeing this coming to fruition certainly pushes me to think of enhancements the QML GUI and it is relatively straightforward to implement the things from QML into PySide, and vice versa.
At the end of the day I would consider the QML interface to get a more feature rich, and eventually contain more than is required for everyday publishing tasks and better suited for in-depth control and customisability, whereas the PySide GUI could potentially take on introductory level and everyday publishing tasks.
Where this goes exactly I'll leave to circumstance. Maybe it'll replace QML. It certainly has greater potential for contribution as many of you already know PySide. On the other hand, by the looks of it, this year or the next will be the year of Qt 5 in which case QML might get easier to develop for and integrate with.
All in all, I'm in favour of alternatives and more minds at the table!