Posts tagged Programming Languages

Random Code 4

:: CodeCritic, Programming Languages

By: John Clements

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
  , slide: function (type, next) {
      var $active = this.$element.find('.active')
        , $next = next || $active[type]()
        , isCycling = this.interval
        , direction = type == 'next' ? 'left' : 'right'
        , fallback  = type == 'next' ? 'first' : 'last'
        , that = this

      this.sliding = true

      isCycling && this.pause()

      $next = $next.length ? $next : this.$element.find('.item')[fallback]()

      if ($next.hasClass('active')) return

      if (!$.support.transition && this.$element.hasClass('slide')) {
        this.$element.trigger('slide')
        $active.removeClass('active')
        $next.addClass('active')
        this.sliding = false
        this.$element.trigger('slid')
      } else {
        $next.addClass(type)
        $next[0].offsetWidth // force reflow
        $active.addClass(direction)
        $next.addClass(direction)
        this.$element.trigger('slide')
        this.$element.one($.support.transition.end, function () {
          $next.removeClass([type, direction].join(' ')).addClass('active')
          $active.removeClass(['active', direction].join(' '))
          that.sliding = false
          setTimeout(function () { that.$element.trigger('slid') }, 0)
        })
      }

random code 3: first look at FORTRAN

:: CodeCritic, Programming Languages

By: John Clements

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
  !================================================================================================================================
  !

  !>Initialises the interpolated point metrics for an interpolated point.
  SUBROUTINE FIELD_INTERPOLATED_POINT_METRICS_INITIALISE(INTERPOLATED_POINT,INTERPOLATED_POINT_METRICS,ERR,ERROR,*)

    !Argument variables
    TYPE(FIELD_INTERPOLATED_POINT_TYPE), POINTER :: INTERPOLATED_POINT !A pointer to the interpolated point to initliase the interpolated point metrics for
    TYPE(FIELD_INTERPOLATED_POINT_METRICS_TYPE), POINTER :: INTERPOLATED_POINT_METRICS !<On exit, a pointer to the interpolated point metrics that have been initialised
    INTEGER(INTG), INTENT(OUT) :: ERR !<The error code
    TYPE(VARYING_STRING), INTENT(OUT) :: ERROR !<The error string
    !Local Variables
    INTEGER(INTG) :: NUMBER_OF_XI_DIMENSIONS,NUMBER_OF_X_DIMENSIONS
    INTEGER(INTG) :: DUMMY_ERR
    TYPE(COORDINATE_SYSTEM_TYPE), POINTER :: COORDINATE_SYSTEM
    TYPE(VARYING_STRING) :: DUMMY_ERROR,LOCAL_ERROR

    CALL ENTERS("FIELD_INTERPOLATED_POINT_METRICS_INITIALISE",ERR,ERROR,*999)

    IF(ASSOCIATED(INTERPOLATED_POINT)) THEN
      IF(ASSOCIATED(INTERPOLATED_POINT_METRICS)) THEN
        CALL FLAG_ERROR("Interpolated point metrics is already associated.",ERR,ERROR,*998)
      ELSE
        NULLIFY(COORDINATE_SYSTEM)
        CALL FIELD_COORDINATE_SYSTEM_GET(INTERPOLATED_POINT%INTERPOLATION_PARAMETERS%FIELD,COORDINATE_SYSTEM,ERR,ERROR,*999)
        NUMBER_OF_X_DIMENSIONS=COORDINATE_SYSTEM%NUMBER_OF_DIMENSIONS
        NUMBER_OF_XI_DIMENSIONS=INTERPOLATED_POINT%INTERPOLATION_PARAMETERS%FIELD%DECOMPOSITION%MESH%NUMBER_OF_DIMENSIONS
        IF(NUMBER_OF_X_DIMENSIONS==SIZE(INTERPOLATED_POINT%VALUES,1)) THEN
          ALLOCATE(INTERPOLATED_POINT_METRICS,STAT=ERR)
          IF(ERR/=0) CALL FLAG_ERROR("Could not allocate interpolated point metrics.",ERR,ERROR,*999)
          ALLOCATE(INTERPOLATED_POINT_METRICS%GL(NUMBER_OF_XI_DIMENSIONS,NUMBER_OF_XI_DIMENSIONS),STAT=ERR)
          IF(ERR/=0) CALL FLAG_ERROR("Could not allocate interpolated point metrics convariant tensor.",ERR,ERROR,*999)
          ALLOCATE(INTERPOLATED_POINT_METRICS%GU(NUMBER_OF_XI_DIMENSIONS,NUMBER_OF_XI_DIMENSIONS),STAT=ERR)
          IF(ERR/=0) CALL FLAG_ERROR("Could not allocate interpolated point metrics contravariant tensor.",ERR,ERROR,*999)
          ALLOCATE(INTERPOLATED_POINT_METRICS%DX_DXI(NUMBER_OF_X_DIMENSIONS,NUMBER_OF_XI_DIMENSIONS),STAT=ERR)
          IF(ERR/=0) CALL FLAG_ERROR("Could not allocate interpolated point metrics dX_dXi.",ERR,ERROR,*999)
          ALLOCATE(INTERPOLATED_POINT_METRICS%DXI_DX(NUMBER_OF_XI_DIMENSIONS,NUMBER_OF_X_DIMENSIONS),STAT=ERR)
          IF(ERR/=0) CALL FLAG_ERROR("Could not allocate interpolated point metrics dXi_dX.",ERR,ERROR,*999)
          INTERPOLATED_POINT_METRICS%INTERPOLATED_POINT=>INTERPOLATED_POINT
          INTERPOLATED_POINT_METRICS%NUMBER_OF_X_DIMENSIONS=NUMBER_OF_X_DIMENSIONS
          INTERPOLATED_POINT_METRICS%NUMBER_OF_XI_DIMENSIONS=NUMBER_OF_XI_DIMENSIONS
          INTERPOLATED_POINT_METRICS%GL=0.0_DP
          INTERPOLATED_POINT_METRICS%GU=0.0_DP
          INTERPOLATED_POINT_METRICS%DX_DXI=0.0_DP
          INTERPOLATED_POINT_METRICS%DXI_DX=0.0_DP
          INTERPOLATED_POINT_METRICS%JACOBIAN=0.0_DP
          INTERPOLATED_POINT_METRICS%JACOBIAN_TYPE=0
        ELSE
          LOCAL_ERROR="The number of coordinate dimensions ("//TRIM(NUMBER_TO_VSTRING(NUMBER_OF_X_DIMENSIONS,"*",ERR,ERROR))// &
            & ") does not match the number of components of the interpolated point ("// &
            & TRIM(NUMBER_TO_VSTRING(SIZE(INTERPOLATED_POINT%VALUES,1),"*",ERR,ERROR))//")."
          CALL FLAG_ERROR(LOCAL_ERROR,ERR,ERROR,*998)
        ENDIF
      ENDIF
    ELSE
      CALL FLAG_ERROR("Interpolation point is not associated.",ERR,ERROR,*998)
    ENDIF

    CALL EXITS("FIELD_INTERPOLATED_POINT_METRICS_INITIALISE")
    RETURN
999 CALL FIELD_INTERPOLATED_POINT_METRICS_FINALISE(INTERPOLATED_POINT_METRICS,DUMMY_ERR,DUMMY_ERROR,*998)
998 CALL ERRORS("FIELD_INTERPOLATED_POINT_METRICS_INITIALISE",ERR,ERROR)
    CALL EXITS("FIELD_INTERPOLATED_POINT_METRICS_INITIALISE")
    RETURN 1
  END SUBROUTINE FIELD_INTERPOLATED_POINT_METRICS_INITIALISE

Ever taken a look at FORTRAN? Actually, I haven’t, either. FORTRAN is basically the oldest non-assembly language around, which is to say the oldest language where someone actually considered “what would programmers like to say?”, rather than “what should the machine be able to do?”

Also, this is FORTRAN 90, which is a much more recent update of FORTRAN. After a quick reading of the Wikipedia page, it looks like f90 (I’m just tired of writing in all caps) is the first update since 1977, which seems ancient but is in fact about halfway between the first invention of FORTRAN and the present day.

It removes many of the restrictions due to machine limitations (allowing free-form source code), and also allows recursive calls, something that the original FORTRAN didn’t—life is easy when you can statically allocate your stack frames!

It also has modules, and… gee whiz: dynamic allocation? Interesting: sounds like it would be much harder to compile f90 code into super-fast assembly than f77, but I suppose you’d have to ask an expert.

Hmmm… I wonder if there’s a formal semantics for more recent versions of Fortran 90? A quick google search suggests the answer is “no”. On the other hand, I’m not sure who would be interested in such a thing; you might be able to sell it as “this formal semantics could identify bugs in your optimization routines”. Anyone want to pay for this?

Okay, back to the source code. What’s going on here?

The first and most obvious thing here is that F90 clearly has a tradition of LONG_CAPITALIZED_VARIABLE_NAMES. COBOL had this going on too, and also an undergraduate classmate of mine named Celeste (RIP, no joke).

I personally find this fairly distracting; it means that a line of source code can’t contain more than two or three tokens, and means that computations have to be split over a whole bunch of lines. On the other hand, local variables apparently have to be declared at the top, so it’s hard to split a large computation into small ones by giving good names to intermediate values.

I think there’s a fairly large split in the world between people who believe that variables should be declared at the top of functions and those who believe that they should be able to appear anywhere; I think that the decls-only-at-top people are imperative people who are planning on doing all kinds of horrible mutation everywhere, and declaring things at the top is their way of preparing themselves for the horror below. The decls-anywhere people (can you tell which I am?), on the other hand, don’t mutate their bindings; these bindings are just a way of giving meaningful names to parts of computation, and there’s no need to hoist them all to the top.

So F90 has the pull toward decls-anywhere that stems from long identifier names, but it’s used by super-imperative people. How do they get any work done? Okay, now I’m just being snide. Sorry. I should point out here that Fortran is clearly one of the massive success stories of language design; it’s one of the first languages, and still in hugely widespread use, thanks in large part to the total 90-degree turn taken by the scientific computing community into C, which is simultaneously the fastest and least optimizable/maintainable language. See Fran Allen’s comments here! Sorry, off topic.

Other, more minor comments: the error-checking code following the dynamic allocation looks like it could easily be abstracted away, especially considering that the error messages aren’t parameterized over the context in any way that I can see.

Broadly speaking, I think I would make local declarations legal anywhere, and shorten up those variable names. Otherwise, looks nice!

Code Snippet 2: This time in Java

:: CodeCritic, Programming Languages

By: John Clements

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
/**************************************************************
 * 
 * Licensed to the Apache Software Foundation (ASF) under one
 * or more contributor license agreements.  See the NOTICE file
 * distributed with this work for additional information
 * regarding copyright ownership.  The ASF licenses this file
 * to you under the Apache License, Version 2.0 (the
 * "License"); you may not use this file except in compliance
 * with the License.  You may obtain a copy of the License at
 * 
 *   http://www.apache.org/licenses/LICENSE-2.0
 * 
 * Unless required by applicable law or agreed to in writing,
 * software distributed under the License is distributed on an
 * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
 * KIND, either express or implied.  See the License for the
 * specific language governing permissions and limitations
 * under the License.
 * 
 *************************************************************/



package org.openoffice.xmerge.merger;

/**
 *  This is an interface used by the {@link
 *  org.openoffice.xmerge.merger.DiffAlgorithm
 *  DiffAlgorithm} and {@link
 *  org.openoffice.xmerge.merger.MergeAlgorithm
 *  MergeAlgorithm} to access a <code>Document</code>.
 *
 *  @author smak
 */
public interface Iterator {


    /**
     *  Move to next element in the sequence.
     *
     *  @return  The <code>Object</code> of the next element in the sequence.
     *           If there is no next element, then return null.
     */
    public Object next();


    /**
     *  Move to previous element in the sequence.
     *
     *  @return  The <code>Object</code> of the previous element in the sequence.
     *           If there is no previous element, then return null.
     */
    public Object previous();


    /**
     * Move to the beginning of the sequence.
     *
     * @return  The <code>Object</code> of the first element in the sequence.
     *          If it is empty, then return null.
     */
    public Object start();


    /**
     * Move to the end of the sequence.
     *
     * @return  The <code>Object</code> of the last element in the sequence.
     *          If it is empty, then return null.
     */
    public Object end();


    /**
     * Return the current element <code>Object</code> content.
     *
     * @return  The <code>Object</code> at current position.
     */
    public Object currentElement();


    /**
     * Return the total element count in the sequence.
     *
     * @return  The total element count.
     */
    public int elementCount();


    /**
     *  A method to allow the difference algorithm to test whether the
     *  <code>obj1</code> and <code>obj2</code> in the
     *  <code>Iterator</code> are considered equal.  As not every
     *  <code>Object</code> in the <code>Iterator</code> can implement its
     *  own equal method, with this equivalent method, we can allow
     *  flexibility for the <code>Iterator</code> to choose a custom way
     *  to compare two objects.  Two objects can even be compared based on
     *  the position in the <code>Iterator</code> rather than by 
     *  the content via this option.
     *
     *  @param  obj1  The first <code>Object</code>.
     *  @param  obj2  The second <code>Object</code>.
     *
     *  @return  true if equal, false otherwise.
     */
    public boolean equivalent(Object obj1, Object obj2);


    /**
     *  <p>A method to force the <code>Iterator</code> to transverse the tree
     *  again to refresh the content.</p>
     *
     *  <p>It is used mainly for <code>Iterator</code> objects which take a snap
     *  shot instead of dynamically transversing the tree.  The current
     *  position will be set to the beginning.</p>
     */
    public void refresh();
}

Here’s another one, randomly chosen from the source for Apache OpenOffice’s 800K lines of Java (I should point out that there are something like 10M lines of C++ in the project).

A few seconds makes it clear that this is … the iterator interface. Minus the horrible “remove” call, and plus a bunch of other ones, generally fairly inoffensive.

Honestly, I can’t say too many bad things about this file: it’s clear and useful; this is the kind of file you’d like to see when you go looking for information. I had hoped for something more… complex? But the random number generator does what it wants. Darn random number generator!

One point here is that I probably shouldn’t have excluded .h files from my random C file excerpt; in Java, interface files aren’t distinguished from “source” files in the same way they are in C.

Code Snippet 1: A Random File From the Linux Source

:: CodeCritic, Programming Languages

By: John Clements

Questions for the audience:

  • What is this code?
  • How does it work?
  • Could it be cleaner?
  • What parts of the code are easy to understand, and what parts are hard?
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
/*
 * Copyright (C) 2008 Freescale Semiconductor, Inc.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * as published by the Free Software Foundation; either version
 * 2 of the License, or (at your option) any later version.
 */

#include <linux/device.h>	/* devres_*(), devm_ioremap_release() */
#include <linux/gfp.h>
#include <linux/io.h>		/* ioremap_prot() */
#include <linux/export.h>	/* EXPORT_SYMBOL() */

/**
 * devm_ioremap_prot - Managed ioremap_prot()
 * @dev: Generic device to remap IO address for
 * @offset: BUS offset to map
 * @size: Size of map
 * @flags: Page flags
 *
 * Managed ioremap_prot().  Map is automatically unmapped on driver
 * detach.
 */
void __iomem *devm_ioremap_prot(struct device *dev, resource_size_t offset,
				 size_t size, unsigned long flags)
{
	void __iomem **ptr, *addr;

	ptr = devres_alloc(devm_ioremap_release, sizeof(*ptr), GFP_KERNEL);
	if (!ptr)
		return NULL;

	addr = ioremap_prot(offset, size, flags);
	if (addr) {
		*ptr = addr;
		devres_add(dev, ptr);
	} else
		devres_free(ptr);

	return addr;
}
EXPORT_SYMBOL(devm_ioremap_prot);