Re: Can MI be supported? (Was: Is the Writing on the Wall for Ada?)

["Robert I. Eachus" <rieachus@attbi.com>]

Warren W. Gay VE3WWG wrote:

> So why not add "has a" test? Then you could test if instance V "has a"
> X, Y or Z component. You could also say that V is "not is a" X, Y or Z,
> since clearly it is not any one of them by themselves.

I don't know that I would call it "has a" but that is exactly what 
mix-ins do.  But I have wondered whether we got mix-ins too right in Ada 
95.  I just saw a post by Matt Heaney on access discriminants.  There 
are some roles for which the mix-in needs access to the whole object, 
and his particular example (persistant types) is one of them.  But for 
most roles where a role only needs access to its own state variables, 
mix-ins do the job nicely and much more elegantly than access 
discriminants.  But some programmers miss this easy way because it is 
too easy and too elegant.

In the example I gave (a complete executable example) the mix-ins added 
state variables of type Color and Size to a type, with Get and Set 
operations for each.  I did this deliberately because some people worry 
about view conflicts with multiple inheritance.  But this is Ada!  The 
operations overload each other, and potentially additional operations of 
the same name on the object type, but the strong typing means that 
Set(Some_Object, Red) and Set(Some_Object, Small) will not conflict.

You can even, and I have done this, instantiate the same mix-in twice on 
the same base type.  I was implementing two-dimensional sparse arrays as 
linked lists.  One mix-in added row links, one added column links, and 
if I wanted to add a (non-zero) element to a sparse matrix, I said:

          Add(Entry, Some_Matrix, Row_Number);
          Add(Entry, Some_Matrix, Column_Number);

And overload resolution took care of the rest. Since Rows and Columns 
were separate types derived from Integer, I couldn't even try to mix 
them up.  This may sound like I was doing linear algebra on the sparse 
arrays, and to some extent I was.  But the entries were actually records 
with about 160 bytes of data, and I was doing row and column permutation 
operations and vector operations where this representation was perfect. 
  The linked lists didn't have to be in any particular order.




-- 
"As far as I'm concerned, war always means failure." -- Jacques Chirac, 
President of France
"As far as France is concerned, you're right." -- Rush Limbaugh