[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]
excess abstract classes, wrong protocols, etc.
- To: <xtech>
- Subject: excess abstract classes, wrong protocols, etc.
- From: Eric Dean Tribble <tribble>
- Date: Wed, 19 Sep 90 11:48:21 PDT
I just ran into SetStepper: an abstract class that defines no
semantics or protocol different from its superclass. Is there any
reason for this to exist?
In the process I noticed something that I've seen show up in lots of
places: deferred messages with a perfectly good abstract
implementation. the isSubsetOf: message can be implemented in ScruSet
to iterate over all the elements of one set looking for those elements
in the other. This simplifies the basic protocol necessary to
implement another set class.
Finally I found various message categories completely misnamed:
isSubsetOf:, isEmpty, and hasMember: go in testing, not accessing.
Is there any reason for me to not change these?