Dennis Lee Bieber wrote: > On 19 Jul 2006 22:38:17 -0700, "mystilleef" <[EMAIL PROTECTED]> > declaimed the following in comp.lang.python: > > >>permitted should be between an object and its mediator. Messages are >>passed through the system via signals or events or established >>protocols. What are the benefits? Well I can change the implementation >>of any object at will without breaking the whole system. Or I can even > > > Python permits that using properties... But one doesn't have to code > the properties from the get-go... Development can be done with direct > attribute access, and if it is determined that some sort of conditioning > logic is needed, it can be implemented WITHOUT CHANGING THE API. > > This IS an "established protocol" in Python. > >>High coupling is bad. Exposing critical states of an object is bad. Any >>decent computer science book will tell you that. >> > > In Python, using, if needed, properties, this is transparent... The > user of the object doesn't know, or need to know, if it is state or > behavior -- all it sees is that it supplies a value for something, > and/or retrieves a value for something. > >>Obviously if I want to control access to an attribute, it means I don't >>want it altered or I want to perform pre/post conditional computations >>before/after the attribute is accessed. There's a reason it is called >>access control. >> > > Python properties support this, no need to have the user of the > object explicitly call a getter or setter... > > > >>changing states you never bothered to hide to begin with, then the >>wisdom behind careful upfront planning begins to make sense. >> > > If it isn't part of the documented API, it is the user's fault. That > applies to all software...
I'm this thread represents proof tht you can lead a horse to water but you can't make him drink - especially when he thinks the water is Kool-Aid, and that attribute APIs can be abused in ways that method-based ones can't. regards Steve -- Steve Holden +44 150 684 7255 +1 800 494 3119 Holden Web LLC/Ltd http://www.holdenweb.com Skype: holdenweb http://holdenweb.blogspot.com Recent Ramblings http://del.icio.us/steve.holden -- http://mail.python.org/mailman/listinfo/python-list
