LISTSERV mailing list manager LISTSERV 16.0

Help for TEAM-ADA Archives


TEAM-ADA Archives

TEAM-ADA Archives


TEAM-ADA@LISTSERV.ACM.ORG


View:

Message:

[

First

|

Previous

|

Next

|

Last

]

By Topic:

[

First

|

Previous

|

Next

|

Last

]

By Author:

[

First

|

Previous

|

Next

|

Last

]

Font:

Proportional Font

LISTSERV Archives

LISTSERV Archives

TEAM-ADA Home

TEAM-ADA Home

TEAM-ADA  December 2000

TEAM-ADA December 2000

Subject:

Re: sorting code

From:

Steven Deller <[log in to unmask]>

Reply-To:

Steven Deller <[log in to unmask]>

Date:

Mon, 18 Dec 2000 11:45:38 -0500

Content-Type:

text/plain

Parts/Attachments:

Parts/Attachments

text/plain (273 lines)

Jesse,
Both VADS and APEX provide *generic* versions of Quicksort, Heapsort and
Insertion sort in a package named "ordering".  I wrote these about 12 years
ago and have used them successfully in numerous situations.  The spec for
the package discusses tradeoffs, so I have copied it below.

You really need to do some analysis before picking the algorithm.

Are your objects large?
-----------------------
If they are very large then SWAPs will dominate COMPAREs and you need to
pick an algorithm that minimizes SWAPs. It may even pay to build an
"indirect" array, sort that, and then do only the MOVEs required.

What is the *guaranteed* maximum size?
--------------------------------------
If there is no limit, that excludes certain algorithms.  If there is a fixed
limit, then some things can be done to make the algorithm faster, including
converting SWAPs to single MOVEs.

What is the stack or local space size restrictions?
---------------------------------------------------
If you have limited stack then before using Quicksort, you need to make sure
you have enough stack for when the algorithm encounters *worst* case inputs.
Quicksort is recursive and can go to a depth of O(N**2) if you are not
careful.

What is the *expected* sorting size?
------------------------------------
If it is quite small, then you need to minimize OVERHEAD for your
algorithms, possibly at the expense of more COMPAREs or even more SWAPs.

Must you have reproducible times, i.e. times  independent of the initial
ordering?
----------------------------------------------------------------------------
------
If so, then Quicksort is *NOT* appropriate.  Quicksort has an *average*
O(NlnN) performance, but there are input patterns that can produce O(N**2)
and all values in between.  The original Quicksort algorithm was worst case
for pre-SORTED data.

Will the list have *equal* values?
----------------------------------
If the compare may have equal values, and you must have a "stable" sort,
then Insertion sort or some other O(N**2) algorithm is required.  A stable
sort keeps the order of equal items.  It is used when the list has already
been sorted by some other process or algorithm and it is important that the
original order not be disturbed by the new sort.


If you are sure the lists will never be more than 20 long, you need to
seriously look at the OVERHEAD for each algorithm.  I'd suggest NOT using
Quicksort, since it only averages a speed of O(NlnN), and with only a few
entries, it is quite possible to have an input pattern that produces
O(N**2).

I've had good luck with Heapsort, which is guaranteed to be O(NlnN), but
that does not minimize MOVEs so you might want to try an "indirect" Heapsort
if your objects are of any significant size.

An "indirect" sort creates an array with the indices to be sorted.  You then
sort the indices using them as indices into the original array.  Finally you
walk the index array to determine where objects are to be located (you walk
it in a way that only requires one MOVE of the pointed to objects and only
for those that are "mislocated" -- that in itself is an interesting little
algorithm.

Regards,
Steve
[log in to unmask]

-- UNIT: package spec ORDERING
-- FILES: ordering_s.a in verdixlib
--        related file is ordering_b.a in verdixlib
-- PURPOSE: provide ordering functions, including sorting and permuting
-- DESCRIPTION: includes packages QUICKSORT, HEAPSORT, INSERTIONSORT,
--              and PERMUTE.
--              Usage: these packages are instantiated with the Element type
--                     of elements to be sorted, the Index discrete type
that
--                     indexes them, and the Boolean function on the
elements
--                     that orders them.
----------------------------------------------------------------------------
--
--
-- package ORDERING
--
-- Provides several sorting algorithms and a permuting algorithm, which are
-- generic with respect to
--   1. Element type in a List to be sorted, an arbitrary type.
--   2. Index type for indexing into the List, a discrete type.
--      A type List is defined from the Element type and the Index type.
--   3. A relation (Boolean result function) between Elements of the List.
--      For an ascending sort, the relation should be True whenever the
--      first Element is strictly less than the second Element.  For a
--      descending sort, the relation should be True whenever the first
--      Element is strictly greater than the second Element.
-- Only "in-place" algorithms are provided, i.e. algorithms whose space
-- requirement is N or ( N + log_2(N)*e ), where e is some small value.  The
-- log_2(N)*e additional space is taken from the stack.
--
-- The following sorting algorithms are provided, and an indication is given
-- about -- when the algorithm should be chosen:
-- 1. Quicksort (median-of-three): least average sort time of all, but
--    there are data organizations for which sorting time is O(N**2).  the
--    median-of-three approach is used for the split partition value, so
that
--    sorted data is NOT a worst case data organization (unlike the basic
--    quicksort algorithm).
-- 2. Heapsort: guaranteed O(NlnN) time.  Note that the "heap" in this
--    algorithm applies to the type of data structure applied to the
--    list to be sorted; there is NO allocation of space from the heap.
-- 3. Insertion sort: one of the simplest sorts.  It is O(N**2) in time
--    complexity, but has the one advantage that it is stable, i.e. elements
--    with equal values retain their order in the list.
--
-- The following sorting algorithms are NOT provided.  An indication of
-- why the algorithms not provided is given.
-- 1. Shellsort (diminishing increments): This algorithm is consistently
--    bettered by Heapsort, and usually bettered by Quicksort.  Either of
--    those should be used instead.
-- 2. Listmerge sort: This algorithm requires a special list that has a
--    link field in each element that can hold an index value.  Its major
--    advantage is that it is a stable sort in O(NlnN) in time, i.e. it
--    makes a reasonable time-space tradeoff to achieve a stable sort.
--    However, its space requirements violate the "general element",
--    "in-place" sort requirements defined for this package.  A reasonable
--    alternative to providing a stable sort in (NlnN) time is to add
--    a single field to each initial element, initialized to the value
--    of the element's position in the array, and make the "<" function
--    include a compare of this field if the original compared fields
--    are equal.  Thus equal elements are forced non-equal, and the order
--    of equal elements in the array is maintained.
-- 3. Bubblesort.  This algorithm takes O(N**2) time, and has a higher
--    constant factor than insertion sort.
-- 4. Selectionsort.  This algorithm takes O(N**2), and is not stable
--    when done in-place with exchanges, unless considerable complications
--    are added to the algorithm.
--
-- It is helpful to organize the sorting algorithms to aid understanding.
-- The following taxonomy is based on the article:
--   An Inverted Taxonomy of Sorting Algorithms, Susan M. Merritt,
--   Communications of the ACM, Vol. 28, No. 1, January 1985, pp. 96-99.
--
-- All sorting algorithms are based on the abstraction "split into parts,
-- sort the parts, join the sorted parts".  Specific sorting algorithms are
-- conveniently classified as hard-split/easy-join or easy-split/hard-join,
-- based on whether the "work" is done at the start or end of the algorithm.
-- The abstract algorithm is clearly recursive.  Usual programming style
-- converts the recursions to iterations for speed.
--
-- If we independently specify the splitting criteria then we get the
-- following table:
--
--   Splitting criteria        hard-split/easy-join
easy-split/hard-join
--   ------------------        --------------------      -------------------
-
--   Equal-sized parts         Quicksort                 Merge Sort
--   One part is one element   Selection Sort            Insertion Sort
--   One element, in place     Bubble Sort               Sinking Sort
--
-- The Heapsort algorithm is viewed as a Selection Sort that minimizes
-- the comparisons to find the next element by using a data structure
-- (a heap).  The Shell Sort algorithm is viewed as a version of the
-- Insertion Sort that operates independently on various subsets of the
-- input data. An additional class of algorithms* may be viewed as
-- splitting criteria that are varying combinations of Quicksort with
-- Bubble Sort.  (Perhaps the combination of Merge Sort and Sinking Sort
-- would give rise to a similar set of algorithms.)
--
-- * A Class of Sorting Algorithms Based on Quicksort, Roger L. Wainwright,
--   Communications of the ACM, Vol. 28, No. 4, April 1985, pp. 396-403.
----------------------------------------------------------------------------
--
--
-- package PERMUTE
--
-- Provides a generic Permute package.  The generic parameters are an
element
-- of any type, an index of discrete type, and an ordering function.  The
-- package defines a list data type and procedure.   The procedure takes a
-- single argument of list type, and rearranges it to the next lexical
-- permutation.  The exception "no_more_permutations" is raised when the
-- input list is lexically the last permutation.
-- The algorithm is derived from that presented in Dijkstra's "A Discipline
of
-- Programming".
--
.......................................................................... -
-

package ORDERING is

    -- hard-split / easy-join sorting algorithms

    generic
        type Element is private;
        type Index is (<>);
        with function "<" ( left, right: Element ) return Boolean is <>;
    package QuickSort is
        type List is array (Index range <>) of Element;
        procedure QuickSort ( L: in out List );
    end QuickSort;
    pragma Share_Body ( QuickSort, False ) ;

    generic
        type Element is private;
        type Index is (<>);
        with function "<" ( left, right: Element ) return Boolean is <>;
    package HeapSort is
        type List is array (Index range <>) of Element;
        procedure HeapSort ( L: in out List );
    end HeapSort;
    pragma Share_Body ( HeapSort, False ) ;

    -- easy-split / hard-join sorting algorithms

    generic
        type Element is private;
        type Index is (<>);
        with function "<" ( left, right: Element ) return Boolean is <>;
    package InsertionSort is
        type List is array (Index range <>) of Element;
        procedure InsertionSort ( L: in out List );
    end InsertionSort;
    pragma Share_Body ( InsertionSort, False ) ;

    -- permute list to next lexical permutation algorithm

    generic
        type Element is private;
        type Index is (<>);
        with function "<=" (left, right: Element) return Boolean is <>;
    package Permute is
        type List is array (Index range <>) of Element;
        procedure Permute ( L: in out List );
        No_more_permutations : exception ;
    end Permute ;
    pragma Share_Body ( Permute , False ) ;

end ORDERING;

--
.......................................................................... -
-
--
-- DISTRIBUTION AND COPYRIGHT:
--
-- This software is copyright 1985 by the VERDIX Corporation.
-- All rights reserved.  No part of the material protected by
-- this copyright notice may be reproduced or utilized in any
-- form without written permission of the copyright owner.
--
-- DISCLAIMER:
--
-- This software and its documentation are provided "AS IS" and
-- without any expressed or implied warranties whatsoever.
-- No warranties as to performance, merchantability, or fitness
-- for a particular purpose exist.
--
-- Because of the diversity of conditions and hardware under
-- which this software may be used, no warranty of fitness for
-- a particular purpose is offered.  The user is advised to
-- test the software thoroughly before relying on it.  The user
-- must assume the entire risk and liability of using this
-- software.
--
-- In no event shall any person or organization of people be
-- held responsible for any direct, indirect, consequential
-- or inconsequential damages or lost profits.

Top of Message | Previous Page | Permalink

Advanced Options


Options

Log In

Log In

Get Password

Get Password


Search Archives

Search Archives


Subscribe or Unsubscribe

Subscribe or Unsubscribe


Archives

December 2017
November 2017
October 2017
September 2017
June 2017
May 2017
April 2017
January 2017
December 2016
November 2016
October 2016
August 2016
July 2016
June 2016
May 2016
April 2016
March 2016
February 2016
January 2016
December 2015
November 2015
October 2015
September 2015
August 2015
June 2015
May 2015
April 2015
March 2015
February 2015
January 2015
December 2014
November 2014
September 2014
August 2014
July 2014
June 2014
May 2014
April 2014
March 2014
February 2014
January 2014
December 2013
November 2013
October 2013
September 2013
June 2013
May 2013
April 2013
March 2013
February 2013
January 2013
December 2012
November 2012
October 2012
September 2012
August 2012
July 2012
June 2012
May 2012
April 2012
February 2012
January 2012
December 2011
November 2011
October 2011
September 2011
August 2011
July 2011
June 2011
May 2011
April 2011
December 2010
November 2010
October 2010
September 2010
June 2010
May 2010
April 2010
March 2010
February 2010
January 2010
December 2009
November 2009
October 2009
September 2009
August 2009
July 2009
June 2009
May 2009
April 2009
February 2009
January 2009
December 2008
November 2008
October 2008
September 2008
August 2008
June 2008
May 2008
April 2008
March 2008
February 2008
January 2008
December 2007
November 2007
October 2007
September 2007
June 2007
May 2007
March 2007
February 2007
December 2006
November 2006
October 2006
September 2006
August 2006
July 2006
June 2006
May 2006
April 2006
March 2006
February 2006
January 2006
December 2005
November 2005
October 2005
August 2005
July 2005
June 2005
May 2005
April 2005
March 2005
February 2005
January 2005
December 2004
November 2004
October 2004
September 2004
August 2004
July 2004
June 2004
May 2004
March 2004
February 2004
January 2004
December 2003
November 2003
October 2003
September 2003
August 2003
July 2003
June 2003
May 2003
April 2003
March 2003
February 2003
January 2003
December 2002
November 2002
October 2002
September 2002
August 2002
July 2002
June 2002
May 2002
April 2002
March 2002
February 2002
December 2001
November 2001
October 2001
September 2001
August 2001
July 2001
June 2001
May 2001
April 2001
March 2001
February 2001
January 2001
December 2000
November 2000
October 2000
September 2000
August 2000
July 2000
June 2000
May 2000
April 2000
March 2000
February 2000
January 2000
December 1999
November 1999
October 1999
September 1999
August 1999
July 1999
June 1999
May 1999
April 1999
March 1999
February 1999
January 1999
December 1998
November 1998
October 1998
September 1998
August 1998
July 1998
June 1998
May 1998
April 1998
March 1998
February 1998
January 1998
December 1997
November 1997
October 1997
September 1997
August 1997
July 1997
June 1997
May 1997
April 1997
March 1997
February 1997
January 1997
December 1996
November 1996
October 1996

ATOM RSS1 RSS2



LISTSERV.ACM.ORG

Secured by F-Secure Anti-Virus CataList Email List Search Powered by the LISTSERV Email List Manager