private

John Roth johnroth at ameritech.net
Sun Jun 30 00:28:38 CEST 2002


"James Kew" <james.kew at btinternet.com> wrote in message
news:afl3e4$f8abc$1 at ID-71831.news.dfncis.de...
> "Peter Hansen" <peter at engcorp.com> wrote in message
> news:3D1DE132.4A890D9D at engcorp.com...
> > Rhymes wrote:
> > >
> > > Is there any possibility to build up a _real_ private attribute?
> > > Such as C++ private data members...
> >
> > You realize that it's quite possible to get at private data members
> > in C++, don't you?
>
> I can't help feeling this is overly dismissive!
>
> Yes, of course one can circumvent private in C++. But for me the main
value
> of the public and private specifiers is that they _document_ to the
client
> of the class what it should access and what it should not.
>
> I don't -- in my admittedly very limited experience -- find Python
classes
> as immediately self-documenting. It's not obvious from the class
definition
> what is intended to be called or modfied by the client and what is
internal
> to the implementation.
>
> I suspect I'm still in the crossover period from strictly declarative
C/C++
> and the "anything goes" Pythonic way and that, given time, I'll find
my
> groove and feel comfortable.
>
> > This has been discussed about 479 times in the last decade.
>
> I'll bet. I'll shut up about it now.
>
> James
>
>
>





More information about the Python-list mailing list