Re: Data Structures as ADTs (was: NFA to DFA)

["Theodore E. Dennison" <dennison@escmail.orl.mmc.com>]

Matthew Heaney wrote:
> 
> In article <31987C20.167EB0E7@escmail.orl.mmc.com>, "Theodore E. Dennison"
> <dennison@escmail.orl.mmc.com> wrote:
> 
> >JEAN CLAUDE GALLO wrote:
> >>
> >> Boaz Chow wrote:
> >> >
> >> > do you know how to implement SET adt and the operations in Ada?
> >> >
> >>
> >> generic packages
> >> You may found examples of SETS in "Software Components with Ada"
> >> (G.Booch). He explain theory of Set algorithm and Ada83 generic packages
> >
> >give me exactly what I want with one simple type declaration. This
> >is because Ada defines the boolean operators AND and OR to work on
> >ARRAYS of boolean in a bitwise fashion. So why bother with some fancy
> >generic ADT package (that might have bugs)?
> 
> Well, then, why bother with abstraction at all?  Why even bother programming
> in a language that offers direct support for programmer-defined abstractions?
> 

I suppose I could take this silly argument to the other extreme: 
Pretend you have convinced me and suggest that all built-in types
are useless and that only a moron would use Integer or Natural for
array indexes when he could use a generic ADT named 
Sequential_Bounded_Unmanaged_Incrementable_Mathematical_Intgeral,
but I don't really have the energy today.

It's not like your SPARC chip knows what a boolean array is. A boolean
array IS AN ABSTRACTION. And it is an abstraction that happens to come
with all the operations needed of sets. (Hint: this was NOT by accident)

If it makes you feel better, here's an abstraction for you:

type Set is array (Possibilities) of Boolean;
function Union (Left : in Set; Right : in Set) renames "OR";
function Intersection (Left : in Set; Right : in Set) renames "AND";

Have a blast.

-- 
T.E.D.          
                |  Work - mailto:dennison@escmail.orl.mmc.com  |
                |  Home - mailto:dennison@iag.net              |
                |  URL  - http://www.iag.net/~dennison         |