Adding new Attribute Interfaces to java-api.

[Brent Putman]

On 9/24/13 3:17 PM, Cantor, Scott wrote:
>
> There aren't that many things I feel extremely strongly about when it
> comes to Java, but avoiding classes where interfaces are usable is
> definitely one of them. I cannot stand the lack of multiple inheritance in
> the language, so I prefer to avoid going concrete until the last possible
> moment.

Yeah, that is generally my tendency as well.  There is an argument I
guess that if you can't really meaningfully have multiple impls of
something, then interfaces add unnecessary complexity.  But I suppose
the debate is always going to be where to draw the line.

>
> I agree that it's not likely to be universal. Context for example is one I
> keep trying to find, and then remember that the top level is BaseContext.
> And no, I don't like that, but it doesn't bother me very much because
> those aren't really pluggable in the same sense.

Yeah, it wasn't like that originally.  :-) If you look back at the
history, you'll see that originally there were a bunch of interfaces for
Context, Subcontext and SubcontextContainer, as well as (I think)
OperationContext, and perhaps some other things.  The "context tree" was
in fact a completely different beast than it is today.    (And off-hand,
I think the reason we didn't call the "BaseContext" class "Context" like
the original interface  is that it's an abstract superclass, so the
naming conventions dictated either Base- or Abstract-).

Basically, you have to pick your battles and at that particular time I
didn't have the time or energy to have that one (and Chad *was* the
project lead, after all).