comp.lang.ada
 help / color / mirror / Atom feed
From: Georg Bauhaus <rm.dash-bauhaus@futureapps.de>
Subject: Re: Silly and stupid post-condition or not ?
Date: Thu, 02 Feb 2012 14:20:55 +0100
Date: 2012-02-02T14:20:56+01:00	[thread overview]
Message-ID: <4f2a8db8$0$7613$9b4e6d93@newsspool1.arcor-online.net> (raw)
In-Reply-To: <82k445fu9n.fsf@stephe-leake.org>

On 02.02.12 10:40, Stephen Leake wrote:

> post conditions should be _complete_. If the Ada
> post condition language is not strong enough to be complete, don't use
> it; use natural language comments.

First, there is no completeness in most real program systems
product; it seems honest, then, to not assume we could manage
to even potentially achieve it is Post aspects, IMHO.
Yet, program system products use formal language.  The use
of formal language means best effort, not completeness of
all kinds, I think.

Post aspects can help being productive:

If programmers know something about the possible outcomes
of a subprogram, they can act accordingly. If there is a
specification that says: the result is such that if P1(Result)
then also P2(Result), they can act accordingly.

A postcondition gives programmers something to work with.
Since natural language does not change the correspondence of
postcondition and programmers writing based on assumptions
about the postcondition, formal language addressing some
results still produce an informed state of mind, a better
one, I think.

As another example,

   type Match (Found: Boolean) is record
      case Found is
          when True =>
             Value : Big_Natural;
          when False =>
             null;
      end case;
   end record;

   procedure Goldbach (A, B: Big_Natural; Result: Match)
     with Post =>
     (if A + B < Highest_Known_Sum then
       (if Result.Found then
               Is_Prime (A)
           and Is_Prime (B)
           and Result.Value mod 2 = 0)
     else  -- larger
       (if Result.Found then
	 Result.Value mod 2 = 0));

Post says that whenever there is a result and it is larger
than some state-of-the-art value, go look and see whether
A and B are prime. If not Result.Found, then data would be thrown
away in any case. Thus, precisely because there is some
kind of formal incompleteness here, the Post aspect tells
something about the results that will not get thrown away.





  reply	other threads:[~2012-02-02 13:20 UTC|newest]

Thread overview: 84+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-01-30 23:11 Silly and stupid post‑condition or not ? Yannick Duchêne (Hibou57)
2012-01-31  6:47 ` J-P. Rosen
2012-01-31 18:48   ` Jeffrey Carter
2012-01-31 22:02     ` Yannick Duchêne (Hibou57)
2012-01-31  8:54 ` Dmitry A. Kazakov
2012-01-31  9:35   ` Georg Bauhaus
2012-01-31 10:22     ` Dmitry A. Kazakov
2012-01-31 12:33       ` Georg Bauhaus
2012-01-31 13:52         ` Dmitry A. Kazakov
2012-01-31 15:34           ` Georg Bauhaus
2012-01-31 16:24             ` Dmitry A. Kazakov
2012-01-31 19:44               ` Georg Bauhaus
2012-02-01  8:41                 ` Dmitry A. Kazakov
2012-02-01 10:37                   ` stefan-lucks
2012-02-01 10:51                     ` Yannick Duchêne (Hibou57)
2012-02-01 13:49                     ` Yannick Duchêne (Hibou57)
2012-02-01 13:49                     ` Dmitry A. Kazakov
2012-02-01 16:37                       ` stefan-lucks
2012-02-02  1:50                         ` Silly and stupid post�?'condition " Randy Brukardt
2012-02-02  1:56                           ` Silly and stupid postâ?'condition or not ? Yannick Duchêne (Hibou57)
2012-02-03  2:45                             ` Silly and stupid post�?'condition or not ? Randy Brukardt
2012-02-02  8:25                         ` Silly and stupid post‑condition " Dmitry A. Kazakov
2012-02-02  9:01                           ` stefan-lucks
2012-02-02  9:18                           ` stefan-lucks
2012-02-02 10:04                             ` Dmitry A. Kazakov
2012-01-31 22:08         ` Yannick Duchêne (Hibou57)
2012-01-31 17:28 ` Dmitry A. Kazakov
2012-01-31 22:12   ` Yannick Duchêne (Hibou57)
2012-02-01  8:49     ` Dmitry A. Kazakov
2012-02-01  8:36 ` Stephen Leake
2012-02-01 16:30   ` Silly and stupid post-condition " Adam Beneschan
2012-02-02  9:40     ` Stephen Leake
2012-02-02 13:20       ` Georg Bauhaus [this message]
2012-02-02 13:35         ` Yannick Duchêne (Hibou57)
2012-02-03  3:13       ` Randy Brukardt
2012-02-03  3:33         ` Yannick Duchêne (Hibou57)
2012-02-03  8:12         ` Simon Wright
2012-02-07  2:29           ` BrianG
2012-02-07 10:43             ` Simon Wright
2012-02-08  2:25               ` BrianG
2012-02-07 21:15             ` Robert A Duff
2012-02-03  9:11         ` Dmitry A. Kazakov
2012-02-04  3:27           ` Randy Brukardt
2012-02-04 10:15             ` Dmitry A. Kazakov
2012-02-03 12:25         ` Phil Thornley
2012-02-04  9:30         ` Phil Thornley
2012-02-04 12:02         ` Phil Thornley
2012-02-05  6:18           ` Randy Brukardt
2012-02-05 10:23             ` Phil Thornley
2012-02-05 10:55               ` Yannick Duchêne (Hibou57)
2012-02-05 15:03               ` Robert A Duff
2012-02-05 18:04                 ` Phil Thornley
2012-02-05 21:27                   ` Robert A Duff
2012-02-05 23:09                     ` Phil Thornley
2012-02-07  2:05               ` Randy Brukardt
2012-02-07  9:38                 ` Dmitry A. Kazakov
2012-02-05 11:31             ` Yannick Duchêne (Hibou57)
2012-02-05 14:50             ` Robert A Duff
2012-02-07  2:11               ` Randy Brukardt
2012-02-07  2:34             ` BrianG
2012-02-07  4:38               ` Yannick Duchêne (Hibou57)
2012-02-09  3:10               ` Randy Brukardt
2012-02-04 23:07         ` Stephen Leake
2012-02-05  2:49           ` Yannick Duchêne (Hibou57)
2012-02-05  6:29           ` Randy Brukardt
2012-02-05 11:40             ` Yannick Duchêne (Hibou57)
2012-02-07  1:36               ` Randy Brukardt
2012-02-05 15:16             ` Robert A Duff
2012-02-06  4:56               ` Yannick Duchêne (Hibou57)
2012-02-06 14:39                 ` Robert A Duff
2012-02-06 16:12                   ` Yannick Duchêne (Hibou57)
2012-02-07  1:46               ` Randy Brukardt
2012-02-07 17:24                 ` Robert A Duff
2012-02-03  6:26       ` J-P. Rosen
2012-02-03  9:12         ` Dmitry A. Kazakov
2012-02-03  9:48           ` Yannick Duchêne (Hibou57)
2012-02-03 11:09             ` Dmitry A. Kazakov
2012-02-03 11:40               ` Yannick Duchêne (Hibou57)
2012-02-03 13:18                 ` Dmitry A. Kazakov
2012-02-03 14:14                   ` Yannick Duchêne (Hibou57)
2012-02-03 14:45                     ` Dmitry A. Kazakov
2012-02-04  3:16           ` Randy Brukardt
2012-02-04  6:27             ` Yannick Duchêne (Hibou57)
2012-02-04 10:47             ` Dmitry A. Kazakov
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox