Hi!
So, here's my next question about object orientation in Ada.
In languages like Ruby, C++ or Java, the following construct is
possible (here in Java):
class State {
HashMap<char, State> transitions = new HashMap<char, State>();
}
Perhaps you're wondering, what this code is meant to be. Some time ago
I programmed a stack-structure simulating a trie as part of a
incremental search-algorithm. The idea was to have a state-object
containing a mapping of characters pointing to state-objects again.
So, the result is a data-structure containing instances of its own
data type again.
As object orientation is a bit different in Ada, the previous code
example would be something like this in Ada:
package State is
type State is tagged
record
Transitions : Transition_Maps.Map;
end record;
package Transition_Maps is new Ada.Containers.Ordered_Maps
(Key_Type => Character,
Element_Type => State,
"<" => "<",
"=" => "=");
use Transition_Maps;
end State;
As you can see, we have a "What had been there first? The hen or the
egg?"-problem: The State-record requires an instance of
Transition_Maps.Map which is declared after the State-record's
definition. But because Transition_Maps.Map needs a State as
Element_Type changing the order ends in the same problem.
So, has anybody an idea, how to solve this paradox?
Thanks,
Matthias
|
