On 1/28/02 8:17 AM, "Chris Withers" <[EMAIL PROTECTED]> wrote:

> Jeffrey P Shell wrote:
>> On 1/27/02 11:25 AM, "Steve Alexander" <[EMAIL PROTECTED]> wrote:
>>> Hi folks,
>>> When I define an Interface, are the methods of the interface supposed to
>>> have "self" as the first argument?
>> No.
> Can you expand on this a little?
> It doesn't make sense to me to exclude 'self'...

Jim [Fulton] explained it to me once.  He can explain it better than I can.
But think of it this way -- theoretically, an interface can be implemented
by a folder full of Python scripts as [relatively] easily as it can be
implemented by a class.  Should you then include all of the binding options
(context, container, etc)?

Or, if you were defining the interface in IDL (mmm,
almost-avoiding-redundancy through acronyms!) with the target language being
Python, would you include self?

An interface defines a contract on the usage of an object/component; it's
not necessarily a specification on how to implement it.  The concept of
'self' as the first argument for instance methods is a product of
implementing in a Python class (not to mention that 'self' is not a required
name, just an extremely common convention.  You could use 'this' just as

The big theory behind most component models is that the implementation
shouldn't matter so long as certain interfaces are satisfied.  And the
following simple interface could be satisfied a number of ways in Zope
already - by a normal Python class implementation; by a folder with a Python
Script or External Method; by a ZClass with the same; or (in Python 2.2) by
a class with a static method (which would have no 'self').

class ISubscriber(Base):
    """ Subscriber base interface """
    def EventService_inform(event):
        """ Informs the subscriber that the event has occurred """

Jeffrey P Shell 

Zope-Dev maillist  -  [EMAIL PROTECTED]
**  No cross posts or HTML encoding!  **
(Related lists - 
 http://lists.zope.org/mailman/listinfo/zope )

Reply via email to