Issue 12491: Update glossary documentation for the term 'attribute' (original) (raw)

Issue12491

Created on 2011-07-04 21:24 by orsenthil, last changed 2022-04-11 14:57 by admin. This issue is now closed.

Messages (4)
msg139809 - (view)	Author: Senthil Kumaran (orsenthil) *	Date: 2011-07-04 21:24
Update the term 'attribute' in the glossary http://docs.python.org/dev/glossary.html#term-attribute so that the reader can understand that the term attribute in Python can mean both 'data-attribute' and a 'callable' method.
msg140075 - (view)	Author: Raymond Hettinger (rhettinger) *	Date: 2011-07-10 11:18
The current glossary entry is fine and encompasses was is ordinarily meant by attribute as distinct from a method. We sometimes use the term loosely to mean any value whether callable or not. And sometimes it is used loosely to mean anything that can be looked up with getattr(). Attempts to over-define it will be incorrect for some uses. Also, it is likely to make the glossary entry less understandable.
msg140076 - (view)	Author: R. David Murray (r.david.murray) *	Date: 2011-07-10 12:44
As an experienced Python programmer, I think of 'attribute' as meaning any attribute (method or non-method) of an object or class. I sometimes do use it imprecisely (to my mind) to mean "non-method attribute", and it is usually clear from context what I mean. Raymond, if attribute means only non-method attributes, what is the word for the set of things that contains both method and non-method...ah, attributes...of an object? (I thought we already had this discussion both on python-dev and another issue....) Hmm. Actually looking at the linked entry, it looks correct to me (it covers both method and non-method attributes as far as I can see). It might be clearer if it mentioned that a value can be anything, including a method.
msg140078 - (view)	Author: R. David Murray (r.david.murray) *	Date: 2011-07-10 12:54
OK, I found the other issue and it looks like we agreed to use 'attributes and methods' where the reference was inclusive. I still think that it is less precise to think this way, but it is clearer exposition given the lack of a good term for non-method attributes. So now I agree that we should not change the existing glossary definition of attribute.