From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00 autolearn=ham
	autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 103376,6b6619eb9cada212
X-Google-Attributes: gid103376,public
From: "Matthew Heaney" <matthew_heaney@acm.org>
Subject: Re: Help me to chose between ADA 95 and C++
Date: 1999/12/18
Message-ID: <385bab78_3@news1.prserv.net>
X-Deja-AN: 562340023
Content-transfer-encoding: 7bit
References: <NBBBJNOMKDIAJALCEFIJGELPDGAA.rleif@rleif.com>
   <83b8il$i5k$1@nntp4.atl.mindspring.net> <3859abde_3@news1.prserv.net>
 <83dsno$5v2$1@nntp2.atl.mindspring.net>
Content-Type: text/plain; charset="US-ASCII"
X-Complaints-To: abuse@prserv.net
X-Trace: 18 Dec 1999 15:42:48 GMT, 32.101.8.235
Organization: Global Network Services - Remote Access Mail & News Services
Mime-version: 1.0
Newsgroups: comp.lang.ada
Date: 1999-12-18T00:00:00+00:00
List-Id: <comp.lang.ada>

In article <83dsno$5v2$1@nntp2.atl.mindspring.net> , Richard D Riehle 
<laoXhai@ix.netcom.com>  wrote:

> In general this is true.  From time to time it might be useful to
> expand or contract the exported behavior for a particular numeric
> type.  For example, I saw a design not long ago with this declaration,
>
>        type Counter is private;
>        function Initialize return Counter;
>        function Increment return Counter;
>        function Decrement return Counter;
>
> where the full defintion was a integer type. The designer intended to
> prevent any operations on Counter except increment and decrement. For
> this application, that was a reasonable approach and eliminated any of
> the complications associated with a client doing more arithmetic than
> necessary.

But you can add and even subtract operations from a scalar type:

  type Counter is new Integer;

  function "+" (L, R : Counter) return Counter is abstract;

  function "-" (L, R : Counter) return Counter is abstract;

You can then use Counter'Succ and Counter'Pred.


> I can think of other circumstances where it would be
> useful to hide the actual type definition and export a restricted set
> of operations, or a special set of operations.  Moreover, it might be
> useful to expand the set of exceptions for that type and add pre- and
> post- conditions in the implementation of some operator/operation.

You can also replace operations:

  type Heading_Type_Base delta 1.0/2**(-16) range -720.0 .. 720.0;

  subtype Heading_Type is
    Heading_Type_Base range 0.0 .. Heading_Type_Base'Pred (360.0);

  function Predefined_Add
    (L, R : Heading_Type_Base) return Heading_Type_Base renames "+";

  function "+" (L, R : Heading_Type) return Heading_Type;

  function Predefined_Subtract
    (L, R : Heading_Type_Base) return Heading_Type_Base renames "-";

  function "-" (L, R : Heading_Type) return Heading_Type;

Now you can implement "+" and "-" to have wrap-around behaviour:

  H := 359.0;
  H := H + 2.0;  -- H = 1.0
  H := 5.0;
  H := H - 10.0;  -- H = 355.0


> Suppose, however, that we want to do a little more than that.
> Suppose the Divide operation is modified as (contrived example):
>
>     function "/"(L , R : T) return T is
>         Result : T;
>     begin
>        if R = Zero then ...          -- raise some exception?
>        elsif some-other-condition    -- raise some other exception
>        elsif even-another-condition  -- raise even-another-exception
>        else
>           Result := Division-operation
>        end if;
>        if Result is-in-some condition -- raise post-condition-exception
>            -- potentially more tests
>        else
>           return Result;
>        end if;
>      end "/";
>
> I fully understand the problems with this example, but I also recognize
> that, in the absence of exportable pre- and post-conditions in Ada, we
> have few options but to check for these within the code itself.

But pre- and post-condition checking is a separate issue from overriding
predefined operators of a scalar type.  We could declare your divide op
for a non-private type:

  type T is new Float;

  function Predefined_Divide (L, R : T) return T renames "/";

  function "/" (R, L : T) return T;


You may prefer to not use an infix operator at all:

  type T is new Float;

  function Div (Dividend, Divisor : T) return T;


In any case you're free to raise whatever exceptions are appropriate.

As far as pre- and post-condition checking goes, I prefer to check using
pragma Assert.  No, this is not an official standard, but it is a de
facto standard.  (There's even an AI to add this pragma to the
language.)

The benefit using a pragma is that the assertion check can be turned on
or off using switches, analogous to suppressing constraint checks.


> Under these circumstances, I might want to write my own operators to ensure
> conformity with rules that cannot be described through simple range
> constraints.

Agreed, but this doesn't mean you have to declare the type as private.


> I agree that this is an additional problem.  Deferred constants seem
> to violate many of the principles of object-oriented programming.

They have their place.  If I write an ADT for exclusive use within a
subsystem -- where I know no type derivation will occur -- then I'd
probably use a deferred constant.

However, if you write an ADT for general consumption, and you don't
really know who your clients are, then I'd use a function.  (Again,
because they might do some derivations, and we want the function to be
inherited.  Constants aren't inherited.)

This issue rears its ugly head for another reason.  If you use a
deferred constant, and elaboration of the constant requires nonstatic
initialization, then you won't be able to use pragma Preelaborate:

package Lists is

  pragma Preelaborate;  --illegal...

  type List_Type is private;

  Null_List : List_Type;
  ...
private
  type List_Type is new Controlled with
    record
      Head : Node_Access;
    end record;

  Null_List : constant List_Type := (Controlled with Head => null);
   -- ... because this is nonstatic

end Lists;


We'd like to use the strongest form of categorization pragma possible.
Above, we'd have to change down to pragma Elaborate_Body.

The solution is to declare Null_List as a primitive operation of the
type:

package Lists
  pragma Preelaborate;  -- legal ...
  type List_Type is private;

  function Null_List return List_Type;
    -- ... because this returns its value at run-time

private
  type List_Type is <as above>
end Lists;


> I think we are once again in agreement on most points.

Yes.

--
Creationists attempt to draw a line between evolutionary biology and the
rest of science by remarking that large-scale evolution cannot be
observed.  This tactic fails.  Large-scale evolution is no more
inaccessible to observation than nuclear reactions or the molecular
composition of water.

Abusing Science: The Case Against Creationism
Philip Kitcher