On Mar 5, 3:49 pm, [EMAIL PROTECTED] wrote: > > Classes and modules are really similar. In Python they're really > *really* similar. > > Actually, at this point, that observation may have more of a > subjective component than I'm used to asserting. I pause here for > corroboration and others' perspectives. Aren't they?
If all you use classes for is for the purpose of a namespace, and all of the methods in the class are just staticmethods, then you have used a class to replicate a module, and they aren't just really, *really* similar, they are the pretty much the same (at least BEHAVIORALLY speaking). But you can't do this: import Zmodule zobject = Zmodule() If you are really using the OO model, and creating instances of classes, then you *have* to use a class, a module wont cut it. I'd say, the less you use a class as an instance factory, the more similar that class is to a module. And you can't do this: class Zclass: import Zclass If Zclass is not defined in your local script, you *have* to know what module it is in, as in: from zpackage.zmodule import Zclass I have seen modules compared to classes that implement a Singleton pattern. In this case, a class is being used in an intentionally degenerate way, such that it creates only one instance. When doing this in Python, one has the choice of replicating the standard pattern with a class, or more simply, just using a module. (Python's language/ object model also permits another option, usually referred to as the Borg implementation, in which many Python names bind to many Python instances, but all share the same underlying object state. But this is tangential to your question.) In general I would say the similarity is a behavioral one, an artifact of the language implementation. Modules are for code packaging and namespace definition, classes are for type definition and object modeling. For more discussions about how similar classes are to modules, you might google for "Python singleton". -- Paul -- http://mail.python.org/mailman/listinfo/python-list
