Re: What is the difference and relation between an interface and
its method?
Mark Space wrote:
If you are working with a designer who insists on using interfaces
for all interactions, that's a good thing. There are other options,
but generally using interfaces wherever possible is preferred.
Lew wrote:
Mark Space wrote:
Huh. I'm kind of surprised. Surely anything can be taken to excess,
but I assume that creating interfaces and coding (and testing) to those
interfaces rather than concrete objects wouldn't be controversial at all.
I wonder what design methodologies would make coding to interfaces
unnecessary.
Don't change horses in midstream.
The part that was controversial was "a designer who insists on using
interfaces for all interactions, that's a good thing", not a designer who
"creat[es] interfaces and cod[es] (and test[s]) to those interfaces rather
than concrete objects". Of course I agree that the latter is a good thing,
and I don't think it is controversial. Nor do I think serious thinkers would
propose that "coding to interfaces [is] unnecessary."
Isn't that called a straw-man argument, where one restates someone's position
into a refutable one that was not the original point?
What is controversial is the carrying of that to extremes: one who *insists*
that *all* interactions go through interfaces, and the notion that that
insistence is "a good thing". In this very newsgroup, you may have noticed
that very controversy raging. So yes, it is controversial. The evidence is
the people vehemently disagreeing in public on that very point.
It has come up in the thread "Composition vs. inheritance". There's some talk
there about whether concrete inheritance is evil.
--
Lew