Candidate GUI Libraries

PyQt5

Python-binding of the well-known Qt (v5) library.

Advantages

• Active project
• Widely used
• Good looking, native look and feel
• matplotlib adapter

Disadvantages

• It's very easy to crash the Python interpreter if the Qt API is not used correctly. See
  • PyQt. How to shoot yourself in the foot.
  • What are good practices for avoiding crashes / hangs in PyQt?
• On Windows, such crashes doesn't even print a stack trace nor can they be debugged, because the actual Qt code is native.

Kivy

• see http://kivy.org/#home
• see http://kivy.org/docs/gettingstarted/intro.html

Advantages

• Modern look and feel, multi-touch capable
• All platforms, including iOS and Android(!)
• Very active project
• matplotlib backend available (with limitations), see https://github.com/kivy-garden/garden.matplotlib

Disadvantages

• Bulky widgets, designed to be touched, made for mobiles and tablets. Seems not well suited to design complex UIs. See http://kivy.org/docs/examples/gallery.html
• No real Desktop features like main and context menues, tool bars, tool tips.
• Mobile-like look and feel (this can be a pro and a con).
• Packaging of the final applications seems not to be an easily task.

wxPython

See http://www.wxpython.org/. Not considered for DeDop GUI because it is supported for Python 2.x only.

Advantages

• Native look and feel and Windows, Max OS X

Disadvantages

• Only for Python 2.x. However, a wxPython clone, called Phoenix (http://wiki.wxpython.org/ProjectPhoenix) is under development and also targets Python 3.x.

Tkinter

The standard Python interface to the Tk GUI toolkit. See https://docs.python.org/3.4/library/tkinter.html. Not considered for DeDop GUI because its look and feel is hopelessly outdated.

Advantages

• Bundled with every Python distribution

Disadvantages

• It's so ugly!