This improved appearance has been a nice thing to see, but looks were never really the reason I decided to leave Tkinter. It's much more the fundamental issues of Tk, plus the fact that Tkinter was never really "completed" (not a comment on the original author--there's only so much one person can do) that has made me decide to leave Tk behind.

As an example of the worst Tk issues (at least for what I was doing):

1) Marks screw up when they come to occupy the position. For example, let's say { symbolizes a text mark with right gravity, and } a text mark with left gravity. Remember that marks all have names, so they can all be referred to uniquely. Now construct some text like this: {here} is {some}  marked {text}. Delete all the text. The marks still exist. Now, try to recreate the original text by inserting the appropriate words at the appropriate marks. It won't work for two reasons, the worst of which is that marks that come to occupy the same index in the text can actually change their relative order.

2) There are no abstract references; everything is referred to by its string name. This causes serious problems with name clashes and code abstraction. Tkinter hides some of this, but not enough.

3) Window nesting is completely determined by window naming, which only compounds the problem mentioned in (2).

I don't really mean to diss Tk, I just get so frustrated when I think of what it could have been...

Thanks,
Ken
On 13-Oct-05, at 4:02 PM, Mark Roseman wrote:

People should probably be more aware of work that has been going on with 

Tk recently (after a fairly long hiatus) in terms of greatly improving 

its appearance, adding theming, and more.  It's worth checking out, and 

of course, if there's a way to help get some of these changes into 

Tkinter, the Tk folks would I'm sure be helpful. 


Here's a portion of a note that Jeff Hobbs posted to the Ruby group 

recently, on a similar topic:




-- 
http://mail.python.org/mailman/listinfo/python-list

Reply via email to