1. When dealing with layering, I'm generally more concerned with the layers of abstraction, and how often a user finds themselves crossing those layers. If each layer is insular, then it's okay. But it's a problem if errors often expose the underlying system (especially if early errors cause errors at a later time, that arise from deep in the system). Also if you frequently have to cross layers to do what you need to do, e.g., if you have to use wxPython to make a real application in Wax. Worse if you have to interface directly with wxWindows because of a missing piece in wxPython. These sorts of things make the layers cumbersome, because they don't simplify anything, they just add yet more stuff to understand and more interactions to worry about.

    Efficiency... eh. In a good framework it's usually not a problem, because you are doing what the application developer would have had to do anyway, only you're doing it better.
      posted by Ian Bicking at 12:30:24 AM on September 16, 2004  
  2. Well, Wax does expose part of the underlying wxPython layer. Often very directly. Many Wax classes derive from wxPython classes (e.g. a Wax Button derives from wx.Button), so Wax controls have the same methods as their wxPython counterparts, plus extras.

    There are some areas where it's difficult (or maybe even impossible) to replace wxPython objects with Wax ones. Event objects, for example. And widget.GetFont() returns a wx.Font rather than a Wax Font. I'm hoping to find a reasonable solution for this.
      posted by Hans Nowak at 12:35:19 PM on September 16, 2004  
  3. Hans, you wrote "I have great plans for Wax, but it's uncertain if I will be able to fulfill them." I occasionally work with wxPython, and have long desired a more Pythonic means of working with it. Wax has looked like it will eventually be the answer.

    In the meantime, though, what are your plans, and why the uncertainty? We might be able to help realize those plans, or alleviate the uncertainty.
      posted by Peter Herndon at 11:54:51 AM on September 17, 2004  
  4. I don't think the uncertainty can be alleviated, I'll just have to wait until next month. Then I'll know more and, depending on the outcome, will present a "road map" for Wax.
      posted by Hans Nowak at 01:16:01 PM on September 17, 2004  
  5. Good enough, then, and may your outcome be for the best.
      posted by Peter Herndon at 02:37:31 PM on September 17, 2004