On Tue, Jun 8, 2010 at 9:29 PM, Steven D'Aprano <steve-remove-t...@cybersource.com.au> wrote: > On Tue, 08 Jun 2010 18:12:05 -0700, geremy condra wrote: > >> I didn't argue that Tcl is bad. I argued that a dependency on it is bad >> for python. Would you argue that Python should ship with Perl and Java >> because there are best-of-breed tools in those languages and python >> could leverage that? Of course not. > > Surely that depends on the tool, how difficult it would be for Python to > duplicate the functionality, and how important it is to provide the tool > as a standard product.
Well, I guess that's where we more or less differ. If I felt like there were a piece of infrastructure for my language so critical that I simply had to drag java along to have it, I would seriously question whether it was worth my time to pursue the project at all. > In any case, Python doesn't ship with Tcl and Tk. They are dependencies > *only if you use Tkinter*. It's not compulsory. <snip> Tkinter- its not only capable of making terrible-looking UIs, its also capable of not being there when you need it! This is not a point in its favor. > Now, we might argue that the Python standard library "must" have a GUI > toolkit, in which case it's going to have some non-trivial dependency. I > don't see why it's so much worse to depend on Tcl/Tk compared to some > other external toolkit. I don't argue that. I don't think it or any other GUI toolkit belong in the stdlib, although if I had to pick one I'm pretty sure Tkinter wouldn't be it. > And even if it were undesirable to rely on any external toolkit, I don't > think it's terribly likely that with the resources available to the PSF > anyone is going to create Yet Another GUI Toolkit specifically for > Python. We might agree that, in a perfect world, it would be nice if > Python had no external dependencies at all (well, apart from the OS of > course), but it isn't going to happen anytime soon. Not unless you're > volunteering? :) Actually, I've had to do pretty much this on two separate occasions, and its not fun. Part of the reason why I like the idea of a sumo distribution- you can keep the core trim and low-dependency, while making sure that platforms where bits are cheap have the goodies that developers want. Geremy Condra -- http://mail.python.org/mailman/listinfo/python-list