On Tue, Aug 01, 2006 at 07:29:51PM -0700, Talin wrote: > tomer filiba wrote: > > that's surly anachronism :) > > > > o.__class__ is a little more typing and will surely scare newbies. > > moreover, type(x) and x.__class__ can return different things > > (you can fool __class__, but not type()). > > > > for my part, i'm fine with any form that makes a distinction between > > the metaclass "type" and the inquire-type "type". > > call it o.__class__, gettype() or typeof(), just don't mix that with > > the metaclass > > From a code style perspective, I've always felt that the magical > __underscore__ names should not be referred to ouside of the class > implementing those names. The double underscores are an indication that > this method or property is in most normal use cases referred to > implicitly by use rather than explicitly by name; Thus str() invokes > __str__ and so on.
The paired double underscores indicate that the function is special to the instance's class. C++ converts understand this just fine until you mention that classes are themselves instances at which point the grey matter takes a while to settle again [guilty]. After that reshuffling you are again assaulted because the stack stops. The class of a class is a type but the class of a class of a class is still a type. Turtles all the way down. See the recent thread on python-checkins for some discussion on why "isinstance(ob, type(type))" isn't just legal -- it's backwards compatible! -Jack _______________________________________________ Python-3000 mailing list Python-3000@python.org http://mail.python.org/mailman/listinfo/python-3000 Unsubscribe: http://mail.python.org/mailman/options/python-3000/archive%40mail-archive.com
