Re: Vocabulary matter: Component vs Element vs Item

[Simon Wright <simon@pushface.org>]

"Yannick Duchêne (Hibou57)" <yannick_duchene@yahoo.fr> writes:

> Le Thu, 25 Jul 2013 21:01:38 +0200, Simon Wright <simon@pushface.org>
> a écrit:
>
>> "Yannick Duchêne (Hibou57)" <yannick_duchene@yahoo.fr> writes:
>>
>>> All three words "component", "element" and "item" could be seen as
>>> interchangeable in plain English (or am I wrong?).
>>
>> I have a feeling that in English (BrE) "component" and "element" are
>> more-or-less synonymous, but "item" isn't. "The solution to the
>> problem has five Xs" - X could be component or element, but not item;
>> an item is a thing on its own, whereas a component or an element is
>> one of many.
>
> Sounds correct.
>
>> So I think that the Ada Containers match this notion: a container can
>> have elements (of element_type, eeew), and you replace one of them by
>> an item.
>>
>> I see that the Booch Components have this the other way round:
>>
>>    procedure Insert (C : in out Abstract_Collection; Elem : Item)
>>    is abstract;
>
> After your note above, I now less feel it's an issue to mix both in a
> same domain (or package, to speak Ada). Remains the case you point
> with the swapped choice made in the Booch Components.
>
> May be a way to solve it, is to have an Element_Type which is a
> subtype (as a type aliasing) of Item_Type. This would allow to have
> both mixed case while having consistency between an argument name and
> its type (there are cases where it's better to simply name a thing
> after its type).

I don't think anyone should take the design choices made in the BCs in
1998 as being in any way definitive. I didn't choose Item/Elem, and I
suspect that a lot more effort went into Ada.Containers naming
conventions (they were at least reviewed by the ARG!).

And the BCs are definitely in maintenance now!