File: ch06.html

package info (click to toggle)
texlive-doc 2005.dfsg.2-1
links: PTS
area: main
in suites: etch, etch-m68k
size: 125,100 kB
ctags: 12,558
sloc: xml: 24,267; perl: 14,394; makefile: 727; sh: 324; lisp: 276; java: 159; sed: 4
file content (3381 lines) | stat: -rw-r--r-- 133,241 bytes
parent folder | download | duplicates (2)
<?xml version="1.0" encoding="iso-8859-1"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">

<html xmlns="http://www.w3.org/1999/xhtml">
  <head>
    <meta name="generator"
    content="HTML Tidy for Linux/x86 (vers 1st March 2002), see www.w3.org" />
    <meta http-equiv="Content-Type"
    content="text/html; charset=ISO-8859-1" />

    <title>Chapter&#160;6.&#160;Pictures and Figures</title>
    <link rel="stylesheet" href="mtw.css" type="text/css" />
    <meta name="generator"
    content="DocBook XSL Stylesheets V1.53.0" />
    <link rel="home" href="index.html" title="Making TeX Work" />
    <link rel="up" href="pt02.html"
    title="Part&#160;II.&#160;Elements of a Complex Document" />
    <link rel="previous" href="ch05.html"
    title="Chapter&#160;5.&#160;Fonts" />
    <link rel="next" href="ch07.html"
    title="Chapter&#160;7.&#160;International Considerations" />
  </head>

  <body>
    <div class="navheader">
      <table border="0" cellpadding="0" cellspacing="0"
      width="100%" summary="Navigation table">
        <tr>
          <td align="left">&#160;<a title="Making TeX Work"
          href="index.html"><img src="figures/nav-home.png"
          alt="Home" border="0" /></a>&#160;<a
          title="Chapter&#160;5.&#160;Fonts" href="ch05.html"><img
          src="figures/nav-prev.png" alt="Prev" border="0" />
          </a>&#160;<a
          title="Part&#160;II.&#160;Elements of a Complex Document"
          href="pt02.html"><img src="figures/nav-up.png" alt="Up"
          border="0" /></a>&#160;<a
          title="Chapter&#160;7.&#160;International Considerations"
          href="ch07.html"><img src="figures/nav-next.png"
          alt="Next" border="0" /></a></td>

          <td align="right"><i>Making TeX Work</i> Version 1.0.1
          <span class="alpha-version">(<a
          href="co01.html"><em>Alpha</em></a>)</span></td>
        </tr>
      </table>
    </div>

    <div class="chapter">
      <div class="titlepage">
        <div>
          <h2 class="title"><a id="chap.pictures"
          name="chap.pictures"></a>Chapter&#160;6.&#160;Pictures
          and Figures</h2>
        </div>

        <div>
          <p class="releaseinfo">$Revision: 1.1 $</p>
        </div>

        <div>
          <p class="pubdate">$Date: 2002/08/23 14:31:13 $</p>
        </div>
        <hr class="component-separator" />
      </div>

      <p>Pictures<a id="id2884175" class="indexterm"
      name="id2884175"></a> and figures<a id="id2884189"
      class="indexterm" name="id2884189"></a> are an important
      component of many documents. This chapter explores how they
      can be incorporated into your TeX documents. There are many
      ways to include pictures and figures in TeX. The most
      important considerations are the type of image<a
      id="id2855783" class="indexterm" name="id2855783"></a>, the
      type of printer you will be using, what platform you are
      using, and how portable the document must be.</p>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a
            id="sec.devindrev" name="sec.devindrev"></a>Different
            Kinds of Images</h2>
          </div>
        </div>

        <p>Images<a id="id2883926" class="indexterm"
        name="id2883926"></a> come from many, many different
        sources, but they can be divided into two broad classes:
        bitmapped and scalable (or vector). Bitmapped images<a
        id="id2883939" class="indexterm" name="id2883939"></a><a
        id="id2883949" class="indexterm" name="id2883949"></a> are
        produced whenever an image is scanned from a drawing,
        photograph, or other printed work. They are also produced
        by most simple paint programs. Scalable images<a
        id="id2883964" class="indexterm" name="id2883964"></a><a
        id="id2883971" class="indexterm" name="id2883971"></a><a
        id="id2883981" class="indexterm" name="id2883981"></a> are
        produced by some more sophisticated drawing programs, many
        commercial sources, and some other applications. Both
        classes have advantages and disadvantages.</p>

        <p>Files that contain graphic images usually end with an
        extension<a id="id2890375" class="indexterm"
        name="id2890375"></a> that identifies the format of the
        image. The extension is a common and convenient nickname
        for the image format. In this chapter, I refer to graphic
        image formats by their extension (for example, GIF images
        or XBM images) without explanation. I do this partly
        because it is easy, but also because it is the most common
        way of referring to them, and you don't really need to know
        anything about the image formats to use them. If your DVI
        driver understands PCX images, you just need a PCX image;
        you don't have to have a detailed understanding of the
        format (thank goodness). If you want to know more, refer to
        the filename extension glossary in Appendix&#160;<a
        href="apa.html"
        title="Appendix&#160;A.&#160;Filename Extension Summary">Appendix&#160;A</a>,
        which will help you identify each of the formats discussed
        in this chapter.</p>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2890411"
              name="id2890411"></a>Bitmapped Images</h3>
            </div>
          </div>

          <p>Photographs<a id="id2890419" class="indexterm"
          name="id2890419"></a> and images with a lot of subtle
          detail are almost always stored as bitmaps. Scanners
          always produce bitmapped images<a id="id2890430"
          class="indexterm" name="id2890430"></a>. Some bitmap
          images can be converted into scalable formats. High-end
          graphic packages like \product{CorelDRAW}<a
          id="id2890443" class="indexterm" name="id2890443"></a>
          and \product{\idx{Adobe Illustrator}} can do this, but
          most cannot. The exceptions are line drawings and other
          very high contrast images. Bitmaps are also
          &#8220;cheap&#8221; to print. Neither the computer nor
          the printer must do very much work to print a
          black-and-white bitmap image. (Color bitmap images must
          be dithered before they can be printed, but that's a
          separate consideration because it need be done only once,
          not every time the image is printed.) Bitmapped images
          are also easy to convert from one format to another.</p>

          <p>Unfortunately, bitmap images are very
          device-dependent. They are stored as a two-dimensional
          array of dots, which gives them a fixed resolution. A
          $3\times5$-inch bitmap image that prints correctly on
          your 300dpi laser printer will only be a
          $\frac{3}{4}\times1\frac{1}{4}$-inch picture if you make
          your final copy on a 1200dpi photo-typesetter.</p>

          <p>Bitmap images also require considerable memory and
          disk space to store. The $3\times5$ image described above
          requires more than 150Kb of memory (if it is
          uncompressed).</p>

          <p>A final consideration is that bitmap images do not
          scale very well. Enlarging or reducing the image requires
          either removing some dots (causing a loss of detail) or
          inserting extra dots (frequently giving slanted lines a
          very jagged appearance). Rescaling images by exact
          integer amounts (doubling or tripling its size, for
          example, but not making it 2$\frac{1}{2}$ times as large)
          works reasonably well (except for jagged edges). Shaded
          regions, which are composed of a regular pattern of black
          and white dots, are frequently disrupted by irregular
          &#8220;blotching&#8221; if non-integral scaling is
          used.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2884033"
              name="id2884033"></a>Scalable Images</h3>
            </div>
          </div>

          <p>Many graphic images can be represented more compactly
          as a collection of lines, curves, and other discrete
          elements. Images of this type are called <span
          class="emphasis"><em>scalable</em></span> or <span
          class="emphasis"><em>vector</em></span> images<a
          id="id2884054" class="indexterm" name="id2884054"></a><a
          id="id2884061" class="indexterm" name="id2884061"></a>.
          Instead of storing every pixel in a rectangular array,
          vector images store the encoded instructions for
          &#8220;drawing&#8221; the image. This provides a compact
          representation. A circle, no matter how large, can be
          represented with just a few data points: the position of
          the center, the radius, the width and color of the line
          that forms the circle, and the pattern that fills the
          circle. It is also easy to change the size of the image;
          if you halve every measurement, the image is drawn at one
          half the size with minimal loss of detail.</p>

          <p>One drawback of scalable images is that they require
          considerable computational power to render. Every printer
          ultimately prints the page as a large bitmap; the print
          engine has to translate the lines, curves, and fills of a
          scalable image into a bitmap before this is
          possible.<sup>[<a id="id2884093" name="id2884093"
          href="#ftn.id2884093">79</a>]</sup> Previewing scalable
          images requires translating them into bitmaps to display
          them. This can be a noticeably slow process unless you
          have a very fast computer.</p>

          <p>The other significant drawback of scalable images is
          that they are difficult to translate from one format to
          another. For example, to translate a PostScript image<a
          id="id2884111" class="indexterm" name="id2884111"></a>
          into something that can be printed on a non-PostScript
          printer, you have to have a program that understands all
          of the commands in the PostScript file. A bitmap
          conversion, on the other hand, doesn't require any
          understanding of graphics commands; it simply has to know
          how to rearrange the bits in the array. What it boils
          down to is this: it's a lot easier to write a translation
          program to &#8220;reverse the order of all the bits in
          each byte in each row&#8221; than it is to write one that
          can &#8220;draw the bezier curve with these three control
          points using a dashed, light-blue line $\frac{1}{8}$ of
          an inch wide.&#8221;</p>
        </div>
      </div>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a id="id2884144"
            name="id2884144"></a>Device Independence Revisited</h2>
          </div>
        </div>

        <p>Pictures<a id="id2890779" class="indexterm"
        name="id2890779"></a> and figures<a id="id2890791"
        class="indexterm" name="id2890791"></a> are a foreign
        concept to TeX. Remember, TeX cares only about building
        pages out of boxes and glue. TeX's notion of a picture is
        frequently nothing more than &#8220;something special goes
        here (I don't know what) and it's 5 inches wide, 2 inches
        high, and 1 inch deep.&#8221; This is device independence,
        of a sort, but many of the easiest ways to include pictures
        and figures in TeX do rely on features of a specific DVI
        driver or a specific kind of printer. If document
        portability is an issue, consider carefully before you
        select a particular way of including pictures and
        figures.</p>

        <p>On the bright side, most DVI drivers<a id="id2890822"
        class="indexterm" name="id2890822"></a> provide some
        mechanism for incorporating pictures and figures. As long
        as it is possible to convert the images from the format
        originally used to a format that another DVI driver
        understands, document portability can be achieved. For
        example, it is possible to print PostScript figures on a
        non-PostScript device if they are first converted into
        another format with <b>Ghostscript</b> or some other
        PostScript interpreter. Unfortunately, it is usually
        inconvenient to convert pictures from one format to
        another, and some conversions may distort the images a
        little bit (or a lot).</p>
      </div>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a
            id="sec.ptexpic" name="sec.ptexpic"></a>Using Only
            TeX</h2>
          </div>
        </div>

        <p>This section describes picture and figure environments
        that don't use any external programs. Graphics created in
        this way are entirely device-independent and can be printed
        with any DVI driver.</p>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2890867"
              name="id2890867"></a>Plain TeX</h3>
            </div>
          </div>

          <p>Plain TeX<a id="id2890880" class="indexterm"
          name="id2890880"></a><a id="id2890893" class="indexterm"
          name="id2890893"></a><a id="id2890906" class="indexterm"
          name="id2890906"></a> has no built-in provision for
          creating pictures or figures. It is possible to do simple
          diagrams and graphs by writing macros that place
          individual points on the page. Figure&#160;<a
          href="ch06.html#fig.plaintex"
          title="Figure&#160;6.1.&#160;An example diagram in Plain TeX">
          Figure&#160;6.1</a> shows several data points plotted in
          Plain TeX. The input was derived from macros presented in
          Appendix D of <span class="emphasis"><em>The
          TeXbook</em></span>&#160;[<a
          href="bi01.html#kn:texbook">kn:texbook</a>]; it is shown
          in Example&#160;<a href="ch06.html#ex.plaintexm"
          title="Example&#160;6.1.&#160;The Input for the Plain TeX Diagram">
          Example&#160;6.1</a>.</p>

          <div class="figure">
            <a id="fig.plaintex" name="fig.plaintex"></a>

            <p class="title"><b>Figure&#160;6.1.&#160;An example
            diagram in Plain TeX</b></p>
<pre class="screen">
FIXME:
</pre>
          </div>

          <div class="example">
            <a id="ex.plaintexm" name="ex.plaintexm"></a>

            <p class="title"><b>Example&#160;6.1.&#160;The Input
            for the Plain TeX Diagram</b></p>
<pre class="programlisting">
<a type="simple" show="embed" actuate="onLoad"
href="tex.06.01.tex"></a>
</pre>
          </div>

          <p>With special-purpose fonts, it is possible to make
          more complex figures in Plain TeX. However, LaTeX
          provides a <tt>picture</tt> environment which greatly
          simplifies the process.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="sec.pictures.latex"
              name="sec.pictures.latex"></a>LaTeX</h3>
            </div>
          </div>

          <p>The <tt>picture</tt> environment in LaTeX<a
          id="id2891050" class="indexterm" name="id2891050"></a><a
          id="id2891064" class="indexterm" name="id2891064"></a><a
          id="id2891076" class="indexterm" name="id2891076"></a> is
          implemented on top of the kinds of primitive operations
          shown in Example&#160;<a href="ch06.html#ex.plaintexm"
          title="Example&#160;6.1.&#160;The Input for the Plain TeX Diagram">
          Example&#160;6.1</a>. Working in the <tt>picture</tt>
          environment is a lot like working on graph paper: you
          begin by specifing how big the graph paper is and the
          distance between lines on the paper (the lines aren't
          really there; they're just used for reference), and then
          inside the <tt>picture</tt> environment, you put <span
          class="emphasis"><em>picture elements</em></span> down on
          the page at the intersections of the lines on the
          grid.</p>

          <p>LaTeX provides picture elements for text, boxes, solid
          and empty circles, lines, and arrows. Lines and arrows
          can be drawn at angles, but they are formed by taking
          characters from a special set of fonts so there is a
          limited number of angles available. The advantage of
          using special fonts to draw the lines is that it is
          relatively efficient.</p>

          <p>Figure&#160;<a href="ch06.html#fig.latexpic"
          title="Figure&#160;6.2.&#160;A parallelogram in LaTeX">Figure&#160;6.2</a>
          is a simple figure drawn with LaTeX's <tt>picture</tt>
          environment. The LaTeX input for this environment is
          shown in Example&#160;<a href="ch06.html#ex.latexpicsrc"
          title="Example&#160;6.2.&#160;The LaTeX Input for ">Example&#160;6.2</a>.</p>

          <div class="figure">
            <a id="fig.latexpic" name="fig.latexpic"></a>

            <p class="title"><b>Figure&#160;6.2.&#160;A
            parallelogram in LaTeX</b></p>

            <div class="mediaobject">
              <img src="FIXME:" />
            </div>
          </div>

          <div class="example">
            <a id="ex.latexpicsrc" name="ex.latexpicsrc"></a>

            <p class="title"><b>Example&#160;6.2.&#160;The LaTeX
            Input for <a href="ch06.html#fig.latexpic"
            title="Figure&#160;6.2.&#160;A parallelogram in LaTeX">Figure&#160;6.2</a></b></p>
<pre class="programlisting">
<a type="simple" show="embed" actuate="onLoad"
href="tex.06.02.tex"></a>
</pre>
          </div>

          <div class="section">
            <div class="titlepage">
              <div>
                <h4 class="title"><a id="sec.Fig"
                name="sec.Fig"></a>The epic and eepic styles</h4>
              </div>
            </div>

            <p>Constructing diagrams using <tt>picture</tt>
            primitives is very tedious because each element of the
            picture has to be placed individually. The <span
            class="emphasis"><em>epic</em></span> style extends
            LaTeX's picture environment by adding several
            higher-level commands for picture construction. These
            commands allow you to draw solid, dotted, and dashed
            lines with arbitrary slopes, create matrices and grids
            with a single command, join several independently
            placed elements together, and read a list of points
            from an external file.</p>

            <p>The diagrams produced by <tt>epic.sty</tt><a
            id="id2891283" class="indexterm" name="id2891283"></a>
            are still limited by the fonts available to LaTeX. The
            <tt>eepic.sty</tt><a id="id2891298" class="indexterm"
            name="id2891298"></a> extends <tt>epic.sty</tt> by
            using \special commands to construct the more complex
            figure elements. The \special commands are the same as
            those used by <tt>tpic</tt> (see the &#8220;<a
            href="ch06.html#sec.tpic" title="MetaFont">the section
            called &#8220;MetaFont&#8221;</a>&#8221; section in
            this chapter) and are supported by many DVI
            drivers.</p>

            <p>Figure&#160;<a href="ch06.html#fig.epic"
            title="Figure&#160;6.3.&#160;A figure created with (a) epic, and (b) eepic">
            Figure&#160;6.3</a> shows a figure constructed with the
            <span class="emphasis"><em>epic</em></span> macros.
            Figure&#160;<a href="ch06.html#fig.epic"
            title="Figure&#160;6.3.&#160;A figure created with (a) epic, and (b) eepic">
            Figure&#160;6.3</a> (a) uses <tt>epic.sty</tt>, and (b)
            uses <tt>eepic.sty</tt>. Notice that <tt>eepic.sty</tt>
            provides circles of arbitrary size and smooth lines at
            any angle. The end points of the radial lines in this
            figure were calculated by another program and read from
            a data file by <span
            class="emphasis"><em>epic</em></span>. The source for
            these figures is shown in Example&#160;<a
            href="ch06.html#ex.epicsrc"
            title="Example&#160;6.3.&#160;The epic Input for ">Example&#160;6.3</a>
            (both diagrams were created with the same source).</p>

            <div class="example">
              <a id="ex.epicsrc" name="ex.epicsrc"></a>

              <p class="title"><b>Example&#160;6.3.&#160;The epic
              Input for <a href="ch06.html#fig.epic"
              title="Figure&#160;6.3.&#160;A figure created with (a) epic, and (b) eepic">
              Figure&#160;6.3</a></b></p>
<pre class="programlisting">
<a type="simple" show="embed" actuate="onLoad"
href="tex.06.03.tex"></a>
</pre>
            </div>

            <div class="figure">
              <a id="fig.epic" name="fig.epic"></a>

              <p class="title"><b>Figure&#160;6.3.&#160;A figure
              created with (a) epic, and (b) eepic</b></p>

              <div class="mediaobject">
                <img src="FIXME:" />
              </div>
            </div>
          </div>

          <div class="section">
            <div class="titlepage">
              <div>
                <h4 class="title"><a id="id2891474"
                name="id2891474"></a>The bezier style</h4>
              </div>
            </div>

            <p>The <span class="emphasis"><em>bezier</em></span>
            style<a id="id2891487" class="indexterm"
            name="id2891487"></a> allows you to construct curved
            lines in the LaTeX picture environment. An example is
            shown in Figure&#160;<a href="ch06.html#fig.bezier"
            title="Figure&#160;6.4.&#160;Several bezier curves created with the LaTeX bezier style">
            Figure&#160;6.4</a>. Its source is shown in
            Example&#160;<a href="ch06.html#ex.beziersrc"
            title="Example&#160;6.4.&#160;The Input for ">Example&#160;6.4</a>.
            The grid in this example was created with <span
            class="emphasis"><em>epic</em></span> for
            convenience.</p>

            <div class="example">
              <a id="ex.beziersrc" name="ex.beziersrc"></a>

              <p class="title"><b>Example&#160;6.4.&#160;The Input
              for <a href="ch06.html#fig.bezier"
              title="Figure&#160;6.4.&#160;Several bezier curves created with the LaTeX bezier style">
              Figure&#160;6.4</a></b></p>
<pre class="programlisting">
<a type="simple" show="embed" actuate="onLoad"
href="tex.06.04.tex"></a>
</pre>
            </div>

            <div class="figure">
              <a id="fig.bezier" name="fig.bezier"></a>

              <p class="title"><b>Figure&#160;6.4.&#160;Several
              bezier curves created with the LaTeX bezier
              style</b></p>

              <div class="mediaobject">
                <img src="FIXME:" />
              </div>
            </div>
          </div>

          <div class="section">
            <div class="titlepage">
              <div>
                <h4 class="title"><a id="id2891594"
                name="id2891594"></a>Other styles</h4>
              </div>
            </div>

            <p>In addition to these styles, there are several other
            style files for doing particular types of drawing.
            Styles exist for drawing logical circuit diagrams, bar
            charts, trees, and more.</p>
          </div>

          <div class="section">
            <div class="titlepage">
              <div>
                <h4 class="title"><a id="id2891609"
                name="id2891609"></a>Other approaches</h4>
              </div>
            </div>

            <p>Because constructing LaTeX drawing by hand is
            tedious, several programs have been written that allow
            you to construct diagrams visually and then produce the
            appropriate <tt>picture</tt> environments
            automatically. The <b>texcad</b><a id="id2891633"
            class="indexterm" name="id2891633"></a> program
            distributed with emTeX<a id="id2891642"
            class="indexterm" name="id2891642"></a> is one such
            program. A similar program called <b>xtexcad</b><a
            id="id2891659" class="indexterm" name="id2891659"></a>
            runs under X11 on some platforms, and there is a Fig
            translator for the LaTeX <tt>picture</tt> environment.
            (For more information about Fig, see the &#8220;<a
            href="ch06.html#sec.Fig"
            title="The epic and eepic styles">the section called
            &#8220;The epic and eepic styles&#8221;</a>&#8221;
            section later in this chapter.)</p>
          </div>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2891692"
              name="id2891692"></a>PiCTeX</h3>
            </div>
          </div>

          <p>PiCTeX<a id="id2891701" class="indexterm"
          name="id2891701"></a><a id="id2891712" class="indexterm"
          name="id2891712"></a> is a macro package that you can
          load on top of Plain TeX or LaTeX.<sup>[<a id="id2891727"
          name="id2891727" href="#ftn.id2891727">80</a>]</sup> It
          does an amazing job of plotting mathematical functions<a
          id="id2891735" class="indexterm" name="id2891735"></a>
          and two-dimensional graphs directly in TeX. An example of
          PiCTeX (taken from <span class="emphasis"><em>The PiCTeX
          Manual</em></span>&#160;[<a
          href="bi01.html#mw:pictex">mw:pictex</a>]) is shown in
          Figure&#160;<a href="ch06.html#fig.pictex"
          title="Figure&#160;6.5.&#160;Sample diagrams using PiCTeX">
          Figure&#160;6.5</a>. The source for this figure is shown
          in Example&#160;<a href="ch06.html#ex.pictexsrc"
          title="Example&#160;6.5.&#160;PiCTeX Input for ">Example&#160;6.5</a>.</p>

          <div class="figure">
            <a id="fig.pictex" name="fig.pictex"></a>

            <p class="title"><b>Figure&#160;6.5.&#160;Sample
            diagrams using PiCTeX</b></p>

            <div class="mediaobject">
              <img src="FIXME:" />
            </div>
          </div>

          <div class="example">
            <a id="ex.pictexsrc" name="ex.pictexsrc"></a>

            <p class="title"><b>Example&#160;6.5.&#160;PiCTeX Input
            for <a href="ch06.html#fig.pictex"
            title="Figure&#160;6.5.&#160;Sample diagrams using PiCTeX">
            Figure&#160;6.5</a></b></p>
<pre class="programlisting">
<a type="simple" show="embed" actuate="onLoad"
href="tex.06.05.tex"></a>
</pre>
          </div>

          <p>The primary drawback of PiCTeX is that it requires a
          considerable amount of memory to use. Even relatively
          simple looking graphs require a big TeX to plot with
          PiCTeX. PiCTeX also produces very, very large
          <tt>DVI</tt> files (the graphs are drawn by interpreting
          and plotting each pixel individually). This usually
          results in very large output files from the DVI driver,
          and some printers may have difficulty printing your
          documents.</p>

          <p>The PiCTeX macros are freely available, but the manual
          is not. You can purchase the manual directly from the
          author of PiCTeX or from the TeX User's Group.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2891884"
              name="id2891884"></a>XYPic</h3>
            </div>
          </div>

          <p><b>XYPic</b><a id="id2891897" class="indexterm"
          name="id2891897"></a><a id="id2891907" class="indexterm"
          name="id2891907"></a> is another add-on macro package for
          TeX, LaTeX, and other formats. It provides considerably
          more flexibility than the LaTeX picture environment
          without resorting to the resource-expensive strategy of
          PiCTeX.</p>

          <p><b>XYPic</b> provides many more arrows than LaTeX
          (including curved and self-pointing forms), a wider
          variety of dashed and dotted lines, provision for lines
          that bend and go around \linebreak</p>

          <p>other picture elements, and annotations for lines and
          arrows. It seems especially well suited to typesetting
          commutative diagrams, simple state-transition diagrams,
          and complex annotated matrices.</p>

          <p>The syntax used in <b>XYPic</b> diagrams is not as
          straightforward as LaTeX's picture environment. An
          example of a complex <b>XYPic</b> diagram is shown in
          Figure&#160;<a href="ch06.html#fig.xypic"
          title="Figure&#160;6.6.&#160;An XYPic diagram">Figure&#160;6.6</a>.
          The corresponding source is shown in Example&#160;<a
          href="ch06.html#ex.xypicsrc"
          title="Example&#160;6.6.&#160;The XYPic Input for ">Example&#160;6.6</a>.</p>

          <div class="figure">
            <a id="fig.xypic" name="fig.xypic"></a>

            <p class="title"><b>Figure&#160;6.6.&#160;An XYPic
            diagram</b></p>

            <div class="mediaobject">
              <img src="FIXME:" />
            </div>
          </div>

          <div class="example">
            <a id="ex.xypicsrc" name="ex.xypicsrc"></a>

            <p class="title"><b>Example&#160;6.6.&#160;The XYPic
            Input for <a href="ch06.html#fig.xypic"
            title="Figure&#160;6.6.&#160;An XYPic diagram">Figure&#160;6.6</a></b></p>
<pre class="programlisting">
<a type="simple" show="embed" actuate="onLoad"
href="tex.06.06.tex"></a>
</pre>
          </div>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2892057"
              name="id2892057"></a>DraTeX</h3>
            </div>
          </div>

          <p>DraTeX<a id="id2892066" class="indexterm"
          name="id2892066"></a><a id="id2892077" class="indexterm"
          name="id2892077"></a> is a macro package that can be
          loaded on top of Plain TeX or LaTeX and provides many
          sophisticated drawing features. An example is shown in
          Figure&#160;<a href="ch06.html#fig.dratex"
          title="Figure&#160;6.7.&#160;An DraTeX diagram">Figure&#160;6.7</a>.
          The source is in Example&#160;<a
          href="ch06.html#ex.dratexsrc"
          title="Example&#160;6.7.&#160;The DraTeX Input for ">Example&#160;6.7</a>.</p>

          <div class="figure">
            <a id="fig.dratex" name="fig.dratex"></a>

            <p class="title"><b>Figure&#160;6.7.&#160;An DraTeX
            diagram</b></p>

            <div class="mediaobject">
              <img src="FIXME:" />
            </div>
          </div>

          <div class="example">
            <a id="ex.dratexsrc" name="ex.dratexsrc"></a>

            <p class="title"><b>Example&#160;6.7.&#160;The DraTeX
            Input for <a href="ch06.html#fig.dratex"
            title="Figure&#160;6.7.&#160;An DraTeX diagram">Figure&#160;6.7</a></b></p>
<pre class="programlisting">
<a type="simple" show="embed" actuate="onLoad"
href="tex.06.07.tex"></a>
</pre>
          </div>

          <p>DraTeX offers a wide variety of low-level drawing
          commands: straight lines at arbitrary angles, circles of
          any size, bezier curves, rectangular or polar coordinate
          systems, perspective for three-dimensional figures,
          user-definable shading patterns, clipping, repetition,
          and user-definable drawing objects and commands. A
          supporting package, AlDraTeX<a id="id2892187"
          class="indexterm" name="id2892187"></a>, provides
          high-level drawing commands for pie charts, XY charts,
          bar charts, spread diagrams, grid diagrams, trees, and
          diagram annotations (arrows, edges, labels).</p>

          <p>Although it is a flexible and portable solution,
          beware that complex figures may require considerable time
          to compute and will almost certainly require a big TeX.
          The resulting <tt>DVI</tt> files tend to be large as
          well.</p>

          <p>DraTeX and AlDraTeX are described in <span
          class="emphasis"><em>TeX &amp; LaTeX: Drawing and
          Literate Programming</em></span>&#160;[<a
          href="bi01.html#eg:dratex">eg:dratex</a>].</p>
        </div>
      </div>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a id="id2892246"
            name="id2892246"></a>Using a Little Help</h2>
          </div>
        </div>

        <p>On most platforms, if you have TeX, MetaFont<a
        id="id2892255" class="indexterm" name="id2892255"></a> is
        available. Because MetaFont was designed for drawing, it
        has much better support for pictures and figures than TeX.
        There are three utilities that allow you to combine TeX and
        MetaFont to create pictures: <b>MFPic</b> and
        <b>Fig2MF</b>.</p>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2892297"
              name="id2892297"></a>MFPic</h3>
            </div>
          </div>

          <p><b>MFPic</b>'s <tt>picture</tt> environment and
          LaTeX's <tt>picture</tt> environment are implemented very
          differently. The commands in <b>MFPic</b> don't attempt
          to typeset your diagram with special fonts; instead, they
          write MetaFont commands to another file. This file must
          be processed with MetaFont before your document can be
          viewed.</p>

          <p>Because MetaFont is used to draw the actual diagram,
          the <b>MFPic</b> macros are much more flexible than the
          LaTeX picture creation commands. <b>MFPic</b> macros
          provide lines (at any angle), rectangles, polygons,
          circles (of any size), ellipses, cyclic and acyclic
          curves, arcs, and wedges (all empty, shaded, or
          filled).</p>

          <p>Unlike <b>Fig2MF</b>, described below, <b>MFPic</b>
          can plot user-specified functions (parametrically or
          directly, using polygonal or bezier interpolation).</p>

          <p>Finally, because <b>MFPic</b> works so directly with
          TeX, it can include labels and captions in the figure as
          well as allowing &#8220;MetaFont hackers&#8221; to insert
          MetaFont code directly.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2892403"
              name="id2892403"></a>Fig2MF</h3>
            </div>
          </div>

          <p>The &#8220;Fig&#8221; in <b>Fig2MF</b><a
          id="id2892422" class="indexterm" name="id2892422"></a>
          stands for the Fig graphics code (in this case, Fig
          version 2.1). Fig is a device-independent way of
          representing figures. Like TeX <tt>DVI</tt> files, Fig
          graphics<a id="id2892446" class="indexterm"
          name="id2892446"></a> are always translated into a
          device-dependent form before they are printed.
          <b>Fig2MF</b> translates Fig graphics into MetaFont code,
          which can be rendered into a font that is usable by TeX
          on any platform.</p>

          <p>At present, there are a few limitations on
          <b>Fig2MF</b>: text objects are ignored, and arrowed and
          non-solid line styles are not supported. The advantage of
          <b>Fig2MF</b> is that very good interactive drawing
          programs, like <b>xfig</b>, can be used to create
          figures.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="sec.tpic"
              name="sec.tpic"></a>MetaFont</h3>
            </div>
          </div>

          <p>If the idea of programming directly in MetaFont<a
          id="id2892507" class="indexterm" name="id2892507"></a> is
          appealing to you, start by reading &#8220;Simple Drawings
          in MetaFont&#8221;&#160;[<a
          href="bi01.html#zw:simple">zw:simple</a>]. This is a
          short document (freely available) that describes how
          simple drawings can easily be rendered directly in
          MetaFont.</p>

          <p>In order to use MetaFont creatively, <span
          class="emphasis"><em>The \MF{}book</em></span>&#160;[<a
          href="bi01.html#kn:mfbook">kn:mfbook</a>] is really a
          necessity, but you can get a feel for the picture
          creation process in MetaFont without it.</p>
        </div>
      </div>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a id="id2892560"
            name="id2892560"></a>Using a Little More Help</h2>
          </div>
        </div>

        <p>A couple of TeX macro packages combine the convenience
        of using only TeX commands to produce pictures and figures
        with the power of PostScript<a id="id2892571"
        class="indexterm" name="id2892571"></a><a id="id2892581"
        class="indexterm" name="id2892581"></a>. Using these macros
        makes your documents less portable because they rely on a
        PostScript printer or PostScript interpreter to be
        printed.</p>
      </div>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a
            id="sec.pictures.pstricks"
            name="sec.pictures.pstricks"></a>PSTricks</h2>
          </div>
        </div>

        <p>The PSTricks<a id="id2892610" class="indexterm"
        name="id2892610"></a><a id="id2892620" class="indexterm"
        name="id2892620"></a> macro package is a TeX-PostScript
        hybrid. The advantage of this approach is that PSTricks is
        able to provide much wider functionality than the preceding
        packages without leaving TeX. The disadvantage is that it
        makes your documents dependent on a PostScript printer.</p>

        <p>One of the neatest features of PSTricks is the ability
        to interact with other TeX objects. You can add PostScript
        annotations (curved lines, labels, etc.) to the entries in
        a table, for example, or point to other elements on the
        page.</p>

        <p>In addition to allowing you to insert essentially
        arbitrary PostScript commands directly into your document,
        PSTricks provides TeX macros for most common picture
        elements including: commands for curved and straight lines
        (at arbitrary angles), a wide variety of arrow heads and
        tails, pattern-filled regions, regular and irregular
        polygonal shapes, text scaling and rotation, and grids.</p>
      </div>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a id="id2892664"
            name="id2892664"></a>TeXdraw</h2>
          </div>
        </div>

        <p>TeXdraw, like \product{\PSTricks}, provides support for
        pictures and figures by relying on a PostScript back
        end.</p>

        <p>TeXdraw<a id="id2892679" class="indexterm"
        name="id2892679"></a><a id="id2892688" class="indexterm"
        name="id2892688"></a> is organized into a toolbox of simple
        routines from which more complex commands can be
        constructed. The TeXdraw manual&#160;[<a
        href="bi01.html#pk:texdraw">pk:texdraw</a>] includes
        several examples of how complex commands can be built from
        toolbox routines.</p>
      </div>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a id="id2892716"
            name="id2892716"></a>tpic</h2>
          </div>
        </div>

        <p>The &#8220;pic&#8221;<a id="id2892728" class="indexterm"
        name="id2892728"></a> drawing language was designed for
        <b>troff</b> documents. The <b>tpic</b><a id="id2892749"
        class="indexterm" name="id2892749"></a><a id="id2892756"
        class="indexterm" name="id2892756"></a><a id="id2892766"
        class="indexterm" name="id2892766"></a> program interprets
        pic drawings and produces TeX output to render them. The
        output relies on a set of \special commands, which must be
        implemented by the DVI driver to actually produce the
        output. The <b>tpic</b> \specials are implemented by many
        DVI drivers.</p>
      </div>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a
            id="sec.scaleformat" name="sec.scaleformat"></a>Using a
            Lot of Help</h2>
          </div>
        </div>

        <p>Although many nice effects can be achieved directly in
        TeX or with MetaFont, you will probably want to include
        some other form of image (e.g., an encapsulated PostScript
        figure or a scanned photograph) in a TeX document
        eventually.</p>

        <p>There are a myriad of choices when this occurs. To a
        large extent, the options available depend on the kind of
        printer you have, which DVI drivers you use, and to what
        extent you are willing to sacrifice device
        independence.</p>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2892815"
              name="id2892815"></a>Electronic Cut-and-paste</h3>
            </div>
          </div>

          <p>Documents<a id="id2892828" class="indexterm"
          name="id2892828"></a> that require complex graphics may
          be difficult to produce as a single TeX document. You may
          exceed the memory limitations of your version of TeX if
          you try to put too many figures on a given page, or you
          may wish to include graphics from incompatible macro
          packages (suppose you are using the Lollipop format for
          example, and you want to include a LaTeX
          <tt>picture</tt>).</p>

          <p>You can always leave blank space in your document,
          print the graphic on a separate page, and then combine
          the two pages with scissors, glue, and a photocopy
          machine. However, there are many times when this is
          inappropriate (not to mention the aesthetic sensibilities
          it may disturb).</p>

          <p>Another solution is to use a program that can perform
          cut-and-paste operations on the output files before they
          are printed. If you use Plain TeX or LamSTeX, one option
          is <b>DVIpaste</b><a id="id2892877" class="indexterm"
          name="id2892877"></a>. <b>DVIpaste</b> can insert parts
          of one <tt>DVI</tt> file into another. It is part of the
          LamSTeX distribution. The documentation for
          <b>DVIpaste</b> contains a good example of its use.
          Unfortunately, <b>DVIpaste</b> relies on macros that are
          incompatible with most other macro packages.</p>

          <p>Some DVI drivers like <b>dvimsp</b><a id="id2892928"
          class="indexterm" name="id2892928"></a>,<sup>[<a
          id="id2892937" name="id2892937"
          href="#ftn.id2892937">81</a>]</sup> which comes with
          emTeX, can translate <tt>DVI</tt> files into bitmapped
          images that can then be incorporated into your document
          directly. Figures&#160;<a href="ch04.html#fig.caffeine"
          title="Figure&#160;4.11.&#160;Caffeine by ChemTeX">Figure&#160;4.11</a>,
          <a href="ch04.html#fig.lithium"
          title="Figure&#160;4.13.&#160;A lithium cation rendered by ChemStruct">
          Figure&#160;4.13</a>, and &#160;<a
          href="ch04.html#fig.mozart"
          title="Figure&#160;4.15.&#160;A little Mozart&#8230;">Figure&#160;4.15</a>
          in Chapter&#160;<a href="ch04.html"
          title="Chapter&#160;4.&#160;Macro Packages">Chapter&#160;4</a>,
          <span class="emphasis"><em><a href="ch04.html"
          title="Chapter&#160;4.&#160;Macro Packages">Chapter&#160;4</a></em></span>,
          were inserted into this book using <b>dvidot</b><a
          id="id2893022" class="indexterm"
          name="id2893022"></a>.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2893032"
              name="id2893032"></a>Scalable Image Formats</h3>
            </div>
          </div>

          <p>Several of the most commonly encountered scalable
          image formats are PostScript (usually encapsulated)<a
          id="id2893044" class="indexterm" name="id2893044"></a><a
          id="id2893054" class="indexterm" name="id2893054"></a>,
          HPGL (Hewlett-Packard's<a id="id2893063"
          class="indexterm" name="id2893063"></a> Plotter
          language), DXF (Autocad's vector format)<a id="id2893072"
          class="indexterm" name="id2893072"></a>, and the Fig
          graphics<a id="id2893081" class="indexterm"
          name="id2893081"></a> language. PostScript and HPGL are
          used by specific printers and plotters, DXF files are
          created by many computer-aided design programs, and Fig
          is a graphic language understood by several free editing
          tools---it is always converted into something else before
          it is printed. Many programs that can edit scalable
          images have their own vendor-specific formats. Luckily,
          these programs can usually produce PostScript or HPGL
          output as well.</p>

          <p>PostScript images are the most commonly encountered
          scalable images. Drawing packages that output PostScript
          generally produce <span class="emphasis"><em>encapsulated
          PostScript</em></span> (EPS)<a id="id2893109"
          class="indexterm" name="id2893109"></a><a id="id2893117"
          class="indexterm" name="id2893117"></a>. Similarly, most
          PostScript clip-art is distributed in EPS format.
          Encapsulated PostScript is a subset of the PostScript
          language. One of the most important features of EPS
          images is a <span class="emphasis"><em>bounding
          box</em></span><a id="id2893135" class="indexterm"
          name="id2893135"></a><a id="id2893145" class="indexterm"
          name="id2893145"></a>. The bounding box identifies the
          size of the EPS image. This information is used to
          determine the scaling factor required to get an image of
          the correct size.</p>

          <p>Generic PostScript can sometimes be converted into
          encapsulated PostScript. For example, the
          <b>ps2epsi</b><a id="id2893170" class="indexterm"
          name="id2893170"></a> program that comes with
          <b>Ghostscript</b><a id="id2893185" class="indexterm"
          name="id2893185"></a> attempts to convert generic
          PostScript into encapsulated PostScript.</p>

          <p>Some software produces reasonable PostScript figures,
          but fails to include the bounding box. For those
          situations, the <b>bbfig</b><a id="id2893206"
          class="indexterm" name="id2893206"></a> program that
          comes with <b>dvips</b> may help. <b>bbfig</b> uses the
          printer to calculate a bounding box that you can insert
          into the figure. After you have printed the figure, you
          can construct a bounding box by hand, if necessary
          (sometimes <b>bbfig</b> gets confused). Simply measure
          the height and width of the figure and its position on
          the page (measured from the bottom-left corner). The
          bounding box is a rectangle measured from the lower-left
          corner of the figure to the upper-right corner. An
          example is shown in Figure&#160;<a
          href="ch06.html#fig.epsbox"
          title="Figure&#160;6.8.&#160;A PostScript bounding box example">
          Figure&#160;6.8</a>.</p>

          <div class="figure">
            <a id="fig.epsbox" name="fig.epsbox"></a>

            <p class="title"><b>Figure&#160;6.8.&#160;A PostScript
            bounding box example</b></p>

            <div class="mediaobject">
              <img src="FIXME:" />
            </div>
          </div>

          <p>The bounding box is measured in PostScript points.
          There are 72 PostScript points to the inch, so the image
          in Figure&#160;<a href="ch06.html#fig.epsbox"
          title="Figure&#160;6.8.&#160;A PostScript bounding box example">
          Figure&#160;6.8</a> has the following bounding box:</p>

          <div class="informalexample">
            <p>&lt;!-- %BoundingBox: 115.2 180 468 626.4 --&gt;</p>
          </div>

          <p>If you are printing your document on a PostScript
          printer, you will find that PostScript images are easy to
          handle; every PostScript DVI driver that I have seen
          allows you to incorporate encapsulated PostScript with
          one or more \special commands. The section &#8220;<a
          href="ch06.html#sec.incps"
          title="Bitmapped Image Formats">the section called
          &#8220;Bitmapped Image Formats&#8221;</a>&#8221; later in
          this chapter describes two ways of incorporating
          encapsulated PostScript images into Plain TeX or LaTeX
          documents for printing on a PostScript printer.</p>

          <p>PostScript output on non-PostScript devices is much
          more difficult, although it can be achieved with a
          PostScript interpreter like <b>Ghostscript</b>.<sup>[<a
          id="id2893346" name="id2893346"
          href="#ftn.id2893346">82</a>]</sup> For a detailed
          description of how PostScript images can be converted
          into another format with <b>Ghostscript</b>, see the
          &#8220;<a href="ch06.html#sec.gs"
          title="Image Magick">the section called &#8220;Image
          Magick&#8221;</a>&#8221; section later in this
          chapter.</p>

          <p>Hewlett-Packard HPGL<a id="id2893388"
          class="indexterm" name="id2893388"></a> is the plotter
          language developed by HP for its line of pen-based
          plotting devices. Many computer-aided design and drafting
          programs can produce HPGL. In addition to plotters, the
          HP LaserJet III and more recent HP LaserJet printers also
          understand HPGL. On an HP LaserJet III or later model
          printer, you can print HPGL directly if your DVI driver
          allows you to include printer-specific data. This ability
          was introduced in version 1.4t of the emTeX
          <b>dvihplj</b><a id="id2893402" class="indexterm"
          name="id2893402"></a> driver.</p>

          <p>If your DVI driver does not support printer-specific
          files or your printer does not understand HPGL, there are
          at least two conversion programs (<b>hp2xx</b><a
          id="id2893432" class="indexterm" name="id2893432"></a>
          and <b>hp2ps</b><a id="id2893446" class="indexterm"
          name="id2893446"></a>) that may be able to convert your
          HPGL drawings into another form you can use.</p>

          <p>The Fig graphics language was designed to represent
          pictures and figures in an easy to interpret and portable
          manner. The <b>Fig</b><a id="id2893467" class="indexterm"
          name="id2893467"></a> and <b>xfig</b><a id="id2893481"
          class="indexterm" name="id2893481"></a> drawing programs
          work with Fig graphics<a id="id2893490" class="indexterm"
          name="id2893490"></a>. Several other programs, like
          <b>gnuplot</b><a id="id2893505" class="indexterm"
          name="id2893505"></a>, can also produce Fig graphics. See
          the &#8220;<a href="ch06.html#sec.Fig"
          title="The epic and eepic styles">the section called
          &#8220;The epic and eepic styles&#8221;</a>&#8221;
          section for more information.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="sec.incps"
              name="sec.incps"></a>Bitmapped Image Formats</h3>
            </div>
          </div>

          <p>Including bitmapped images<a id="id2893540"
          class="indexterm" name="id2893540"></a> is easier than
          including scalable images because no real interpretation
          of commands is necessary. Over the years, lots of
          different ways have been developed for including
          bitmapped images in TeX documents. Since the development
          of the <b>bm2font</b> program, most of these methods have
          become obsolete. As a result, this section examines only
          a few methods.</p>

          <div class="section">
            <div class="titlepage">
              <div>
                <h4 class="title"><a id="id2893563"
                name="id2893563"></a>DVI driver {\ttbackslash
                special}</h4>
              </div>
            </div>

            <p>The first method is DVI driver-specific: if your DVI
            driver includes a \special command for including
            bitmapped graphic images, you can simply use that
            command. The disadvantage of using a DVI driver
            \special is that it makes the document less portable.
            Instead of being printable with any DVI driver, it now
            requires a DVI driver that recognizes a particular
            \special command. Note also that some DVI drivers do
            not handle color images very well; you may need to
            convert the image to black and white first. (Many
            programs described in the &#8220;<a
            href="ch06.html#sec.picconv"
            title="Manipulating Images">the section called
            &#8220;Manipulating Images&#8221;</a>&#8221; section
            later in this chapter can perform this operation.)</p>
          </div>

          <div class="section">
            <div class="titlepage">
              <div>
                <h4 class="title"><a id="id2893598"
                name="id2893598"></a>bm2font</h4>
              </div>
            </div>

            <p>On most platforms, <b>bm2font</b><a id="id2893612"
            class="indexterm" name="id2893612"></a> provides an
            ideal solution for including bitmapped images. This
            program translates bitmapped images into <tt>PK</tt><a
            id="id2893630" class="indexterm" name="id2893630"></a>
            fonts and produces a snippet of TeX code that can be
            used to print the image in your document. Because most
            DVI drivers can use <tt>PK</tt> fonts, a high degree of
            portability is maintained. Some portability is still
            lost since the bitmapped image has a fixed resolution,
            but that is a consequence of using the bitmapped image,
            not <b>bm2font</b>. All of the bitmap images in this
            book were included with the <b>bm2font</b> program.</p>

            <p>This book is a good example of an instance where
            portability is required. I could easily have included
            the graphics with the \special command of the DVI
            driver that I use most frequently (<b>dvihplj</b>), but
            then I couldn't have produced pages suitable for final
            publication.</p>

            <p><b>bm2font</b> can read a number of common graphic
            image formats including <tt>GIF</tt><a id="id2893714"
            class="indexterm" name="id2893714"></a>, Windows
            <tt>BMP</tt><a id="id2893731" class="indexterm"
            name="id2893731"></a>, and <tt>PCX</tt><a
            id="id2893753" class="indexterm" name="id2893753"></a>.
            It can translate color images into black and white
            using a number of dithering<a id="id2893763"
            class="indexterm" name="id2893763"></a> methods, and it
            can scale the image to a specific size (although
            scaling bitmap images is not usually very
            effective).</p>
          </div>

          <div class="section">
            <div class="titlepage">
              <div>
                <h4 class="title"><a id="id2893776"
                name="id2893776"></a>pbmtopk</h4>
              </div>
            </div>

            <p>The <b>pbmtopk</b><a id="id2893790"
            class="indexterm" name="id2893790"></a> program
            provides a solution similar to <b>bm2font</b>. Because
            <b>pbmtopk</b><a id="id2893812" class="indexterm"
            name="id2893812"></a> is a smaller program, it may be
            easier to port to other systems. The <b>pbmtopk</b>
            distribution includes <b>pktopbm</b><a id="id2893834"
            class="indexterm" name="id2893834"></a>, which can
            translate <tt>PK</tt> files back into <tt>PBM</tt><a
            id="id2893860" class="indexterm" name="id2893860"></a>
            bitmaps.</p>
          </div>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2893872"
              name="id2893872"></a>Inserting PostScript Images into
              TeX</h3>
            </div>
          </div>

          <p>If you are using Plain TeX or LaTeX and printing your
          documents on a PostScript printer<a id="id2893883"
          class="indexterm" name="id2893883"></a>, there are
          several (essentially equivalent) style files that you can
          use to include your figures.</p>

          <p>One is <tt>epsfig.sty</tt><a id="id2893906"
          class="indexterm" name="id2893906"></a>, which can be
          used as a LaTeX style option or be inserted directly into
          Plain TeX documents with \input.<sup>[<a id="id2893914"
          name="id2893914" href="#ftn.id2893914">83</a>]</sup> The
          <tt>epsfig</tt> style option is supported by LaTeX2e as
          well.</p>

          <p>After the style file is loaded, you can use the macro
          \epsfig to include your figure. The complete syntax for
          \epsfig is:</p>
<pre class="screen">
  \epsfig{figure=, height=, width=, rheight=, rwidth=, 
          bbllx=, bblly=, bburx=, bbury=,
          clip=, angle=, silent=}
</pre>

          <p>In practice, only a few of these options are commonly
          used. Here is a description of each:</p>

          <div class="variablelist">
            <dl>
              <dt><span class="term"><tt>figure=</tt></span></dt>

              <dd>
                <p>Identifies the name of the file containing the
                PostScript figure. This option is required.</p>
              </dd>

              <dt><span class="term"><tt>height=,
              width=</tt></span></dt>

              <dd>
                <p>Specify the height and width of the figure. If
                only one is specified, the other will be scaled
                automatically to keep the proportions of the
                original figure. If both are given, the figure will
                be scaled (anamorphically) to the requested
                size.</p>
              </dd>

              <dt><span class="term"><tt>rheight=,
              rwidth=</tt></span></dt>

              <dd>
                <p>Provide the &#8220;reserved&#8221; height and
                width of the figure. This is how big the TeX box
                that encloses the figure will be. By default, the
                box is as big as the figure.</p>
              </dd>

              <dt><span class="term"><tt>bbllx=, bblly=, bburx=,
              bbury=</tt></span></dt>

              <dd>
                <p>Specify the bounding box<a id="id2894103"
                class="indexterm" name="id2894103"></a> of the
                figure. If not specified (it usually isn't), the
                bounding box is read from the PostScript
                figure.</p>
              </dd>

              <dt><span class="term"><tt>clip=</tt></span></dt>

              <dd>
                <p>Indicates whether or not the figure should be
                &#8220;clipped&#8221; at its bounding box. Clipping
                prevents lines in the figure from extending beyond
                the bounding box.</p>
              </dd>

              <dt><span class="term"><tt>angle=</tt></span></dt>

              <dd>
                <p>Allows you to specify that the figure should be
                rotated. Always specify the angle of rotation in
                degrees. The figure is always rotated
                counter-clockwise.</p>
              </dd>

              <dt><span class="term"><tt>silent=</tt></span></dt>

              <dd>
                <p>Turns off informative messages printed by the
                macros as the figure is processed.</p>
              </dd>
            </dl>
          </div>

          <p>The parameters to the \epsfig macro have to obey TeX's
          strict parsing rules. In particular, you must not put
          spaces around the equal sign in any option. The
          <tt>clip=</tt> and <tt>silent=</tt> options have no
          values, but you must include the equal sign anyway.</p>

          <p>For example, Figure&#160;<a
          href="ch06.html#fig.psfigure"
          title="Figure&#160;6.9.&#160;An example of an encapsulated figure">
          Figure&#160;6.9</a> was inserted by \epsfig using the
          following commands:</p>
<pre class="screen">
\begin{figure}
  \epsfig{figure=figs/sample.eps}
  \caption{An example of an encapsulated figure.}
  \label{fig.psfigure}
\end{figure}
</pre>
          <a id="id2894252" class="indexterm" name="id2894252"></a>

          <div class="figure">
            <a id="fig.psfigure" name="fig.psfigure"></a>

            <p class="title"><b>Figure&#160;6.9.&#160;An example of
            an encapsulated figure</b></p>

            <div class="mediaobject">
              <img src="FIXME:" />
            </div>
          </div>

          <p>Figure&#160;<a href="ch06.html#fig.psfigurerot"
          title="Figure&#160;6.10.&#160;Another example of an encapsulated figure (resized and rotated)">
          Figure&#160;6.10</a> demonstrates a few of other
          <tt>epsfig</tt> options:</p>
<pre class="screen">
\begin{figure}
  \epsfig{figure=figs/sample.eps,width=6cm,angle=45}
  \caption{Another example of an encapsulated figure 
           (resized and rotated)}
  \label{fig.psfigurerot}
\end{figure}
</pre>
          <a id="id2894324" class="indexterm" name="id2894324"></a>

          <div class="figure">
            <a id="fig.psfigurerot" name="fig.psfigurerot"></a>

            <p class="title"><b>Figure&#160;6.10.&#160;Another
            example of an encapsulated figure (resized and
            rotated)</b></p>

            <div class="mediaobject">
              <img src="figs/tex.06.09.eps" />
            </div>
          </div>
        </div>
      </div>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a
            id="sec.picconv" name="sec.picconv"></a>Manipulating
            Images</h2>
          </div>
        </div>

        <p>There are literally dozens of software packages
        available for manipulating graphic images<a id="id2894380"
        class="indexterm" name="id2894380"></a>. The discussion
        that follows focuses on a small handful of these programs.
        The programs selected are representative of the kinds of
        tools available, but this is by no means an endorsement
        that these are the best tools. Table&#160;<a
        href="ch06.html#tab.manipover"
        title="Table&#160;6.1.&#160;Graphics Manipulation Packages">
        Table&#160;6.1</a> summarizes the programs described in
        this section.</p>

        <p>If you cannot find any way of converting a particular
        image, remember that you may be able to display the image
        and capture it with another program that can save the image
        in a more tractable format.</p>

        <div class="table">
          <a id="tab.manipover" name="tab.manipover"></a>

          <p class="title"><b>Table&#160;6.1.&#160;Graphics
          Manipulation Packages</b></p>

          <table summary="Graphics Manipulation Packages"
          border="1">
            <colgroup>
              <col align="left" />
              <col align="left" />
              <col align="left" />
              <col align="left" />
            </colgroup>

            <thead>
              <tr>
                <th align="left">\bf Program</th>

                <th align="left">\bf Platforms</th>

                <th align="left">\bf Cost</th>

                <th align="left">\bf Purpose</th>
              </tr>
            </thead>

            <tbody>
              <tr>
                <td align="left">Image Magick</td>

                <td align="left">unix</td>

                <td align="left">Free</td>

                <td align="left">Display, manipulate, convert and
                capture images</td>
              </tr>

              <tr>
                <td align="left">PBMplus</td>

                <td align="left">All</td>

                <td align="left">Free</td>

                <td align="left">Convert and manipulate images</td>
              </tr>

              <tr>
                <td align="left">xv</td>

                <td align="left">unix</td>

                <td align="left">SW${}^1$</td>

                <td align="left">Display, manipulate, convert and
                capture images</td>
              </tr>

              <tr>
                <td align="left">xloadimage</td>

                <td align="left">unix</td>

                <td align="left">Free</td>

                <td align="left">Display images</td>
              </tr>

              <tr>
                <td align="left">Image Alchemy</td>

                <td align="left">MS-DOS${}^2$</td>

                <td align="left">SW</td>

                <td align="left">Convert images</td>
              </tr>

              <tr>
                <td align="left">ColorView</td>

                <td align="left">MS-DOS</td>

                <td align="left">SW</td>

                <td align="left">Convert images</td>
              </tr>

              <tr>
                <td align="left">Jpeg4</td>

                <td align="left">MS-DOS</td>

                <td align="left">SW</td>

                <td align="left">Convert images</td>
              </tr>

              <tr>
                <td align="left">pmjpeg</td>

                <td align="left">OS/2</td>

                <td align="left">SW</td>

                <td align="left">Display and convert images</td>
              </tr>

              <tr>
                <td align="left">txt2pcx</td>

                <td align="left">MS-DOS</td>

                <td align="left">SW</td>

                <td align="left">Translate text into graphic image
                format</td>
              </tr>

              <tr>
                <td align="left">Ghostscript</td>

                <td align="left">All</td>

                <td align="left">Free</td>

                <td align="left">Convert PostScript to other
                formats</td>
              </tr>

              <tr>
                <td align="left">GoScript</td>

                <td align="left">MS-DOS</td>

                <td align="left">\$\$${}^3$</td>

                <td align="left">Convert PostScript to other
                formats</td>
              </tr>

              <tr>
                <td align="left">hp2xx</td>

                <td align="left">unix</td>

                <td align="left">Free</td>

                <td align="left">Convert HPGL to other formats</td>
              </tr>

              <tr>
                <td align="left">\multicolumn{4}{l}{ ${}^1$\vrule
                height11pt width0pt\tiny Shareware: &#8220;Try
                before you buy,&#8221; but not free.}</td>

                <td class="auto-generated">&#160;</td>

                <td class="auto-generated">&#160;</td>

                <td class="auto-generated">&#160;</td>
              </tr>

              <tr>
                <td align="left">\multicolumn{4}{l}{ ${}^2$\tiny
                Also available for Sun workstations; see
                documentation.}</td>

                <td class="auto-generated">&#160;</td>

                <td class="auto-generated">&#160;</td>

                <td class="auto-generated">&#160;</td>
              </tr>
            </tbody>
          </table>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="sec.gs"
              name="sec.gs"></a>Image Magick</h3>
            </div>
          </div>

          <p><b>Image Magick<a id="id2894787" class="indexterm"
          name="id2894787"></a></b> is a collection of eight
          programs for manipulating images. These programs work
          with MIFF files<a id="id2894797" class="indexterm"
          name="id2894797"></a> by default. MIFF is the
          &#8220;machine-independent file format&#8221; developed
          by the package's author. If you are going to manipulate
          images with these tools, the author recommends converting
          them to MIFF format first. The <b>convert</b> program
          reads and writes a large number of image formats, so
          conversion to and from MIFF format is
          straightforward.</p>

          <p><b>Image Magick</b> requires the Independent JPEG
          Group<a id="id2894834" class="indexterm"
          name="id2894834"></a>'s JPEG library in order to
          manipulate JPEG images. Similarly, Sam Leffler<a
          id="id2894844" class="indexterm" name="id2894844"></a>'s
          TIFF software library is required to manipulate TIFF<a
          id="id2894854" class="indexterm" name="id2894854"></a>
          images. Both of these are compile-time options. If you
          did not build <b>Image Magick</b>, you may not be able to
          manipulate JPEG or TIFF images. To manipulate PostScript
          images, <b>Ghostscript</b> must be available.</p>

          <p>The programs described below are the <b>Image
          Magick</b> tools that allow you to manipulate images in
          ways that may be necessary or useful for creating
          printable images. Other tools, like <b>animate</b><a
          id="id2894899" class="indexterm" name="id2894899"></a>,
          included in the <b>Image Magick</b> distribution, aren't
          described here because they have no bearing on
          printability.</p>

          <div class="variablelist">
            <dl>
              <dt><span class="term"><b>display</b></span></dt>

              <dd>
                <p>This program<a id="id2894938" class="indexterm"
                name="id2894938"></a> allows you to preview images
                on an X11 display. The number of colors in the
                image is reduced to match the number of colors of
                your display, if necessary. Program options allow
                you to specify a variety of image-processing
                operations on the image (noise reduction, scaling,
                manipulation of the color map, etc.).</p>

                <p>Conversion to MIFF format, as recommended above,
                isn't necessary for simply displaying the
                image.</p>
              </dd>

              <dt><span class="term"><b>import</b></span></dt>

              <dd>
                <p>This program<a id="id2894981" class="indexterm"
                name="id2894981"></a> allows you to capture visible
                portions of an X11 display. You can capture any
                visible window, the entire display, or any
                rectangular portion of the display. <b>import</b>
                can save the image in any format recognized by
                <b>convert</b>.</p>
              </dd>

              <dt><span class="term"><b>convert</b></span></dt>

              <dd>
                <p>As the name implies, this utility<a
                id="id2895033" class="indexterm"
                name="id2895033"></a> converts images between
                graphics formats. The following standard formats
                are recognized:</p>

                <table class="simplelist" border="0"
                summary="Simple list">
                  <tr>
                    <td>MIFF</td>

                    <td>BMP</td>

                    <td>CMYK</td>

                    <td>EPS</td>

                    <td>FAX</td>

                    <td>GIF</td>

                    <td>IRIS</td>

                    <td>JPEG</td>

                    <td>PICT</td>

                    <td>PNM</td>
                  </tr>

                  <tr>
                    <td>PS</td>

                    <td>RGB</td>

                    <td>RLE</td>

                    <td>SUN</td>

                    <td>TGA</td>

                    <td>TEXT</td>

                    <td>TIFF</td>

                    <td>XBM</td>

                    <td>XWD</td>

                    <td>&#160;</td>
                  </tr>
                </table>

                <p>Additionally, a few less-standard formats are
                recognized. A complete list is available in the
                manual page for <b>convert</b>.</p>
              </dd>

              <dt><span class="term"><b>combine</b></span></dt>

              <dd>
                <p>This program combines<a id="id2895167"
                class="indexterm" name="id2895167"></a> two images
                by blending them together. The result varies
                tremendously, depending on the kind of blending
                used. Simple overlaps, various kinds of cutouts,
                and more complex color blendings are possible.</p>
              </dd>

              <dt><span class="term"><b>montage</b></span></dt>

              <dd>
                <p>Unlike <b>combine</b>, <b>montage</b><a
                id="id2895217" class="indexterm"
                name="id2895217"></a> combines multiple images by
                <span class="emphasis"><em>tiling<a id="id2895229"
                class="indexterm"
                name="id2895229"></a></em></span>. (The images are
                laid next to each other rather than on top of each
                other.) Optionally, the image name can appear below
                each tile.</p>
              </dd>

              <dt><span class="term"><b>mogrify</b></span></dt>

              <dd>
                <p><b>mogrify</b><a id="id2895268"
                class="indexterm" name="id2895268"></a> is used to
                transform an image (or series of images). You can
                specify alternate color maps in a variety of ways,
                as well as dithering<a id="id2895280"
                class="indexterm" name="id2895280"></a> and error
                correcting transformations. The images can also be
                rolled, rotated, rescaled, and sheared. (Shearing<a
                id="id2895290" class="indexterm"
                name="id2895290"></a> transforms a rectangle into a
                parallelogram by shifting the X or Y axis.)</p>
              </dd>
            </dl>
          </div>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2895306"
              name="id2895306"></a>PBMplus</h3>
            </div>
          </div>

          <p>The <b>PBMplus<a id="id2895320" class="indexterm"
          name="id2895320"></a></b> package is a large collection
          of image translation programs. This package was written
          originally for unix, but has since been ported to MS-DOS
          and OS/2.</p>

          <p><b>PBMplus</b> defines three file formats<a
          id="id2895342" class="indexterm" name="id2895342"></a><a
          id="id2895350" class="indexterm" name="id2895350"></a><a
          id="id2895357" class="indexterm"
          name="id2895357"></a>:</p>

          <div class="itemizedlist">
            <ul type="disc">
              <li>
                <p>PBM---The Portable Bitmap format for black and
                white images</p>
              </li>

              <li>
                <p>PGM---The Portable Graymap format for grayscaled
                images</p>
              </li>

              <li>
                <p>PPM---The Portable Pixmap format for color
                images</p>
              </li>
            </ul>
          </div>
          <br />
          <br />

          <p>Translation from one format to another is accomplished
          by translating into the appropriate portable format
          first, and then translating the portable bitmap into the
          destination format.</p>

          <p>The following image formats are supported by the
          PBMPlus tools:</p>

          <table class="simplelist" border="0"
          summary="Simple list">
            <tr>
              <td>Abekas YUV bytes</td>

              <td>MGR bitmaps</td>
            </tr>

            <tr>
              <td>Andrew Toolkit raster objects</td>

              <td>MTV or PRT ray traced images</td>
            </tr>

            <tr>
              <td>Atari Degas .pi1 images</td>

              <td>MacPaint files</td>
            </tr>

            <tr>
              <td>Atari Degas .pi3 files</td>

              <td>Macintosh PICT files</td>
            </tr>

            <tr>
              <td>Atari Spectrum files</td>

              <td>PCX files</td>
            </tr>

            <tr>
              <td>AutoCAD slides</td>

              <td>PostScript &#8220;images&#8221;</td>
            </tr>

            <tr>
              <td>Bennet Yee &#8220;faces&#8221;</td>

              <td>QRT ray tracer files</td>
            </tr>

            <tr>
              <td>CMU window manager bitmaps</td>

              <td>Raw grayscale bytes</td>
            </tr>

            <tr>
              <td>Doodle brush files</td>

              <td>Raw RGB bytes</td>
            </tr>

            <tr>
              <td>FITS files</td>

              <td>Sun icons</td>
            </tr>

            <tr>
              <td>GIF files</td>

              <td>Sun rasterfiles</td>
            </tr>

            <tr>
              <td>Gould scanner files</td>

              <td>TIFF images</td>
            </tr>

            <tr>
              <td>Group 3 faxes</td>

              <td>TrueVision Targa files</td>
            </tr>

            <tr>
              <td>HIPS files</td>

              <td>Usenix FaceSaver(tm) images</td>
            </tr>

            <tr>
              <td>HP PaintJet files</td>

              <td>X10/X11 window dumps</td>
            </tr>

            <tr>
              <td>IFF ILBM files</td>

              <td>X11 pixmaps</td>
            </tr>

            <tr>
              <td>IMG (GEM) images</td>

              <td>XBM files</td>
            </tr>

            <tr>
              <td>Img-whatnot files</td>

              <td>Xim file</td>
            </tr>

            <tr>
              <td>Lisp Machine bitmaps</td>

              <td>&#160;</td>
            </tr>
          </table>

          <p>The <b>PBMplus</b> tools translate the image into the
          appropriate portable format (PBM, PGM, or PPM) depending
          on the nature of the image. Similarly, the tools can all
          work with &#8220;lower&#8221; formats. (PPM tools can
          work with PGM and PBM files, and PGM tools can work with
          PBM files.) For example, there is no
          &#8220;<b>pbmtopcx</b>,&#8221; however the
          <b>ppmtopcx</b><a id="id2895625" class="indexterm"
          name="id2895625"></a> tool will create a black and white
          PCX file if you use a PBM file as the input. Once
          converted into a portable format, the image can be
          translated back into (almost) any of the supported
          formats.</p>

          <p>In addition to image conversion, the PBMPlus tools
          include programs to perform a wide variety of image
          manipulations.</p>

          <table class="simplelist" border="0"
          summary="Simple list">
            <tr>
              <td>Apply Conway's rules of life</td>

              <td>Create a blank image</td>
            </tr>

            <tr>
              <td>Create a mask bitmap</td>

              <td>Reduce/enlarge an image</td>
            </tr>

            <tr>
              <td>Convert text into a bitmap</td>

              <td>Create a UPC bitmap</td>
            </tr>

            <tr>
              <td>Bentleyize a greymap</td>

              <td>Apply edge-detection</td>
            </tr>

            <tr>
              <td>Apply edge-enhancement</td>

              <td>Generate a histogram</td>
            </tr>

            <tr>
              <td>Normalize contrast</td>

              <td>Apply an &#8220;oil-painting&#8221; filter</td>
            </tr>

            <tr>
              <td>Create a &#8220;ramp&#8221;</td>

              <td>Generate textural features</td>
            </tr>

            <tr>
              <td>Perform bitwise arithmetic</td>

              <td>Perform MxN convolution</td>
            </tr>

            <tr>
              <td>Crop an image</td>

              <td>Cut/Paste rectangles</td>
            </tr>

            <tr>
              <td>Color-reduce</td>

              <td>Flip/rotate</td>
            </tr>

            <tr>
              <td>Apply gamma-correction</td>

              <td>Build a visual index</td>
            </tr>

            <tr>
              <td>Invert an image</td>

              <td>Add a border to an image</td>
            </tr>

            <tr>
              <td>Scale an image</td>

              <td>Shear an image</td>
            </tr>

            <tr>
              <td>Smooth an image</td>

              <td>Build a tiled image</td>
            </tr>

            <tr>
              <td>Colorize a graymap</td>

              <td>Dither color images</td>
            </tr>

            <tr>
              <td>Create fractal terrain</td>

              <td>Apply Laplacian-relief filter</td>
            </tr>
          </table>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2895807"
              name="id2895807"></a>xv</h3>
            </div>
          </div>

          <p><b>xv</b><a id="id2895821" class="indexterm"
          name="id2895821"></a> is an interactive image
          viewing/converting tool that runs on X11 servers
          (including MS-DOS implementations of X11 such as
          Desqview/X).</p>

          <p>You can view GIF<a id="id2895837" class="indexterm"
          name="id2895837"></a>, PBM<a id="id2895845"
          class="indexterm" name="id2895845"></a>, XBM<a
          id="id2895854" class="indexterm" name="id2895854"></a>,
          Sun rasterfile<a id="id2895863" class="indexterm"
          name="id2895863"></a>, JPEG<a id="id2895871"
          class="indexterm" name="id2895871"></a>, and TIFF<a
          id="id2895880" class="indexterm" name="id2895880"></a>
          images. You can save any image in one of those formats or
          in encapsulated PostScript. In addition, you can capture
          any visible portion of the X11 server window
          interactively with <b>xv</b>.</p>

          <p>Like most of the tools described here, <b>xv</b>
          includes a number of image manipulation tools (cropping,
          scaling, editing the color map, etc.). If you try to view
          an image that is larger than your display, <b>xv</b>
          automatically scales it to fit. I find that <b>xv</b>'s
          interactive nature makes it easier to use for image
          manipulation than the command-driven tools.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2895932"
              name="id2895932"></a>xloadimage</h3>
            </div>
          </div>

          <p>The <b>xloadimage</b><a id="id2895946"
          class="indexterm" name="id2895946"></a> program is
          another X11 picture display tool. Like <b>xv</b> and
          <b>Image Magick</b>'s <b>display</b>, it has a number of
          picture manipulation options (although they are not
          interactive).</p>

          <p>One advantage of <b>xloadimage</b> is that it does not
          rescale images that are too large to fit within the
          display. Instead, it places them in a scrollable window.
          I find this behavior superior to <b>xv</b>'s solution of
          rescaling a very large image even though you are
          interested only in a small section of it.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2896003"
              name="id2896003"></a>Image Alchemy</h3>
            </div>
          </div>

          <p><b>Image Alchemy</b><a id="id2896018"
          class="indexterm" name="id2896018"></a> is a shareware
          tool for converting graphic images between various
          formats. Although the manual mentions a version for Sun
          workstations, I have only seen the MS-DOS version.</p>

          <p>Note that the unregistered shareware version of this
          program will only convert images which are 640$\times$480
          pixels or smaller.</p>

          <p>The following is a list of the graphic formats
          recognized by version 1.5 of <b>Image Alchemy</b>:</p>

          <table class="simplelist" border="0"
          summary="Simple list">
            <tr>
              <td>ADEX</td>

              <td>PCPAINT/Pictor Page Format</td>
            </tr>

            <tr>
              <td>Autologic</td>

              <td>PCX</td>
            </tr>

            <tr>
              <td>Binary Information Files (BIF)</td>

              <td>Portable BitMap (PBM)</td>
            </tr>

            <tr>
              <td>Encapsulated PostScript (EPS)</td>

              <td>Q0</td>
            </tr>

            <tr>
              <td>Erdas LAN/GIS</td>

              <td>QDV</td>
            </tr>

            <tr>
              <td>Freedom of the Press</td>

              <td>QRT</td>
            </tr>

            <tr>
              <td>GEM VDI Image File</td>

              <td>Scodl</td>
            </tr>

            <tr>
              <td>GIF</td>

              <td>Silicon Graphics Image</td>
            </tr>

            <tr>
              <td>HP Printer Command Language (PCL)</td>

              <td>Stork</td>
            </tr>

            <tr>
              <td>HP Raster Transfer Language (RTL)</td>

              <td>Sun Raster</td>
            </tr>

            <tr>
              <td>HSI JPEG</td>

              <td>TIFF</td>
            </tr>

            <tr>
              <td>HSI Palette</td>

              <td>Targa</td>
            </tr>

            <tr>
              <td>HSI Raw</td>

              <td>Utah Raster Toolkit (RLE)</td>
            </tr>

            <tr>
              <td>IFF/ILBM</td>

              <td>Vivid</td>
            </tr>

            <tr>
              <td>JPEG/JFIF</td>

              <td>Windows Bitmap (BMP)</td>
            </tr>

            <tr>
              <td>Jovian VI</td>

              <td>WordPerfect Graphic File</td>
            </tr>

            <tr>
              <td>Macintosh PICT/PICT2</td>

              <td>XBM</td>
            </tr>

            <tr>
              <td>MTV Ray Tracer</td>

              <td>XWD</td>
            </tr>
          </table>

          <p><b>Image Alchemy</b> can also perform a number of
          image manipulations like rescaling, cropping, and color
          map changes.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2896234"
              name="id2896234"></a>ColorView</h3>
            </div>
          </div>

          <p><b>ColorView<a id="id2896248" class="indexterm"
          name="id2896248"></a></b> is a shareware MS-DOS program
          for displaying and converting graphic images. Although
          designed to work with VESA SuperVGA adapters,
          <b>ColorView</b> can also work with plain VGA adapters.
          It does require a 80286 or a more powerful processor.</p>

          <p><b>ColorView</b> can read JPEG<a id="id2896279"
          class="indexterm" name="id2896279"></a>, GIF (87 and
          89)<a id="id2896288" class="indexterm"
          name="id2896288"></a>, BMP<a id="id2896296"
          class="indexterm" name="id2896296"></a>, and RLE<a
          id="id2896310" class="indexterm" name="id2896310"></a>
          images. It can write GIF (87), 8-bit BMP, 24-bit BMP,
          JPEG, and RLE images. It can perform image translation
          without displaying the images, so a VGA or SuperVGA
          adapter is not required for translation.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2896325"
              name="id2896325"></a>Jpeg4</h3>
            </div>
          </div>

          <p>This<a id="id2896334" class="indexterm"
          name="id2896334"></a> is the Independent JPEG Group<a
          id="id2896343" class="indexterm" name="id2896343"></a>'s
          JPEG Software. The compression tool can convert GIF<a
          id="id2896353" class="indexterm" name="id2896353"></a>,
          PPM<a id="id2896361" class="indexterm"
          name="id2896361"></a>, and Targa (TGA)<a id="id2896370"
          class="indexterm" name="id2896370"></a> images into
          JPEG<a id="id2896379" class="indexterm"
          name="id2896379"></a> format. The decompression tool
          converts JPEG images into GIF, PPM (or PGM<a
          id="id2896389" class="indexterm" name="id2896389"></a>),
          or Targa formats.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2896400"
              name="id2896400"></a>pmjpeg</h3>
            </div>
          </div>

          <p>The <b>pmjpeg</b><a id="id2896415" class="indexterm"
          name="id2896415"></a> program is an OS/2 2.x program that
          can read and display JPEG<a id="id2896424"
          class="indexterm" name="id2896424"></a>, TIFF<a
          id="id2896433" class="indexterm" name="id2896433"></a>,
          Targa<a id="id2896441" class="indexterm"
          name="id2896441"></a>, GIF<a id="id2896450"
          class="indexterm" name="id2896450"></a>, PCX<a
          id="id2896459" class="indexterm" name="id2896459"></a>,
          and Windows<a id="id2896467" class="indexterm"
          name="id2896467"></a> or OS/2 BMP<a id="id2896482"
          class="indexterm" name="id2896482"></a> images. In
          addition, <b>pmjpeg</b> can write images in any of these
          formats except Windows BMP. <b>pmjpeg</b> can also
          capture all or part of the desktop. Some of the other
          features of <b>pmjpeg</b> are color map editing, contrast
          enhancement, image cropping, scaling, rotation, batch
          translation of images into JPEG format, cyclic slideshow
          presentations, and the ability to reduce an image to the
          system palette colors.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2896523"
              name="id2896523"></a>txt2pcx</h3>
            </div>
          </div>

          <p><b>txt2pcx</b><a id="id2896537" class="indexterm"
          name="id2896537"></a> is a memory-resident MS-DOS utility
          that captures text-mode screens as PCX graphic images.
          See the discussion of screen capturing in the &#8220;<a
          href="ch06.html#sec.screendump" title="GoScript">the
          section called &#8220;GoScript&#8221;</a>&#8221; section
          of this chapter for more information about why this is
          sometimes desirable.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2896564"
              name="id2896564"></a>Ghostscript</h3>
            </div>
          </div>

          <p><b>Ghostscript</b><a id="id2896577" class="indexterm"
          name="id2896577"></a> is distributed by the Free Software
          Foundation (FSF)<a id="id2896586" class="indexterm"
          name="id2896586"></a>. It is one of the few freely
          available programs that can convert PostScript images
          into other formats.</p>

          <p><b>Ghostscript</b> is a (mostly) complete PostScript<a
          id="id2896608" class="indexterm" name="id2896608"></a>
          interpreter. Converting an entire page of PostScript into
          another format is straightforward with
          <b>Ghostscript</b>. The following example converts the
          PostScript file <tt>file.ps</tt> into a monochrome GIF<a
          id="id2896637" class="indexterm" name="id2896637"></a>
          image called <tt>image1.gif</tt>:</p>
<pre class="screen">
\$ <span
class="bold"><b>gs -r300 -sDEVICE=gifmono -sOutputFile=image1.gif file.ps</b></span>
 
</pre>
          <br />
          <br />

          <p>The parameter <span
          class="emphasis"><em>-r300</em></span> indicates that
          <b>Ghostscript</b> should generate output at a resolution
          of 300dpi. This controls the size of the bitmap. If you
          plan to print the resulting bitmap, you should specify
          the same resolution as your printer. If <tt>file.ps</tt>
          contains more than one page, you can use a <tt>&lt;!--
          d</tt> in the output file's name to identify how multiple
          pages --&gt; should be handled. For example, if
          <tt>file.ps</tt> contains three pages, the following
          command will extract the first page into
          <tt>image1.gif</tt>, the second page into
          <tt>image2.gif</tt>, and the third page into
          <tt>image3.gif</tt>:<sup>[<a id="id2896732"
          name="id2896732" href="#ftn.id2896732">84</a>]</sup></p>
<pre class="screen">
\$ <span
class="bold"><b>gs -r300 -sDEVICE=gifmono -sOutputFile=image\%d.gif file.ps</b></span>
 
</pre>
          \vskip-\baselineskip <br />
          <br />

          <p>For incorporating images into TeX, it is more common
          to convert an encapsulated PostScript (EPS)<a
          id="id2896761" class="indexterm" name="id2896761"></a><a
          id="id2896769" class="indexterm" name="id2896769"></a><a
          id="id2896779" class="indexterm" name="id2896779"></a>
          image into a bitmapped format than it is to convert
          entire pages. Encapsulated PostScript is described
          earlier in this chapter in the section &#8220;<a
          href="ch06.html#sec.scaleformat"
          title="Using a Lot of Help">the section called
          &#8220;Using a Lot of Help&#8221;</a>.&#8221; It is
          slightly more difficult to convert encapsulated
          PostScript because of the bounding box<a id="id2896806"
          class="indexterm" name="id2896806"></a>. To convert this
          into a bitmap of exactly the right size, you have to tell
          <b>Ghostscript</b> to move the image to the upper-left
          corner of the page (where <b>Ghostscript</b> always
          begins its conversion) and tell it how big the image is
          in pixels.</p>

          <p>The following steps are necessary to produce exactly
          the right size bitmap of the encapsulated PostScript
          figure:</p>

          <div class="orderedlist">
            <ol type="1">
              <li>
                <p>Find the size of the bounding box. Examine the
                PostScript file using a text editor. Find the line
                that begins with <tt>&lt;!-- % BoundingBox:</tt>
                --&gt; followed by four numbers. Those numbers are
                the lower-left x-coordinate (llx), the lower-left
                y-coordinate (lly), the upper-right x-coordinate
                (urx), and the upper-right y-coordinate (ury),
                respectively.</p>
              </li>

              <li>
                <p>Create another file called <tt>trans.ps</tt>
                that contains the single line:</p>
<pre class="screen">
<span class="emphasis"><em>llx</em></span> neg <span
class="emphasis"><em>lly</em></span> neg translate
</pre>
              </li>

              <li>
                <p>Calculate the width of the bounding box:
                $width_{bb} = urx - llx$.</p>
              </li>

              <li>
                <p>The width you have just calculated is the width
                at 72dpi. To convert this to the resolution that
                you will be using, multiply by the resolution and
                divide by 72. For example, if you will be printing
                at 300dpi and the width of the bounding box
                $width_{bb}$ is 216, the width at 300dpi, <span
                class="emphasis"><em>width</em></span>, is $(216
                \times 300) \div 72 = 900$.</p>
              </li>

              <li>
                <p>Calculate the corrected height in an analogous
                manner using $height_{bb} = ury - lly$ as a
                starting point. For example, if $height_{bb} =
                360$, the height you get at 300dpi is 1500.</p>
              </li>

              <li>
                <p>Finally, run <b>Ghostscript</b> using the <span
                class="emphasis"><em>-g</em></span> parameter to
                select the image size ($width \times height$). For
                example, to translate the EPS file <tt>card.ps</tt>
                into the GIF file <tt>card.gif</tt> at 300dpi
                assuming the height and width calculated above,
                run:</p>
<pre class="screen">
\$ <span
class="bold"><b>{gs -r300 -g900x1500 -sDEVICE=gifmono {\bs}
       -sOutputFile=card.gif trans.ps card.ps}</b></span>
 
</pre>
              </li>
            </ol>
          </div>

          <p>The <span class="emphasis"><em>Perl</em></span> script
          shown in Example&#160;<a href="ch06.html#ex.gseps"
          title="Example&#160;6.8.&#160;Converting Encapsulated PostScript to a Bitmap with Ghostscript">
          Example&#160;6.8</a> automates this process.</p>

          <div class="example">
            <a id="ex.gseps" name="ex.gseps"></a>

            <p class="title"><b>Example&#160;6.8.&#160;Converting
            Encapsulated PostScript to a Bitmap with
            Ghostscript</b></p>
<pre class="programlisting">
<a type="simple" show="embed" actuate="onLoad"
href="gs-eps.pl"></a>
</pre>
          </div>

          <p>The primary disadvantage of <b>Ghostscript</b> is that
          the selection of free PostScript fonts is quite limited.
          The only freely available fonts are: IBM Courier,
          Bitstream Courier, Bitstream Charter, Adobe Utopia, URW
          Antiqua, URW Grotesk Bold, Nimbus Roman No9, and Nimbus
          Sans. Unless you have the PostScript sources for other
          fonts used in your figures (<tt>PFA</tt> or <tt>PFB</tt>
          and <tt>AFM</tt> files), <b>Ghostscript</b> will use
          crude approximations of the desired font.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="sec.screendump"
              name="sec.screendump"></a>GoScript</h3>
            </div>
          </div>

          <p><b>GoScript</b><a id="id2897134" class="indexterm"
          name="id2897134"></a> is a commercial PostScript<a
          id="id2897143" class="indexterm" name="id2897143"></a>
          interpreter sold by LaserGo, Inc.<a id="id2897155"
          class="indexterm" name="id2897155"></a> <b>GoScript</b>
          is available for MS-DOS systems only. A special version
          is available for Microsoft Windows.</p>

          <p>The primary advantage of a commercial interpreter is
          that it comes with more high-quality fonts. The
          <b>GoScript</b> program comes with &#8220;clones&#8221;
          of the 13 standard PostScript fonts:</p>

          <table class="simplelist" border="0"
          summary="Simple list">
            <tr>
              <td>Times Roman</td>

              <td>Times Italic</td>
            </tr>

            <tr>
              <td>Times Bold</td>

              <td>Times Bold Italic</td>
            </tr>

            <tr>
              <td>Helvetica</td>

              <td>Helvetica Oblique</td>
            </tr>

            <tr>
              <td>Helvetica Bold</td>

              <td>Helvetica Bold Oblique</td>
            </tr>

            <tr>
              <td>Courier</td>

              <td>Courier Oblique</td>
            </tr>

            <tr>
              <td>Courier Bold</td>

              <td>Courier Bold Oblique</td>
            </tr>

            <tr>
              <td>Symbol</td>

              <td>&#160;</td>
            </tr>
          </table>

          <p>The <b>GoScript Plus</b> program includes clones of
          all 35 standard fonts (the standard 13 plus AvantGarde,
          Bookman, Helvetica-Narrow, New Century Schoolbook,
          Palatino in four styles, Zapf Chancery Italic, and Zapf
          Dingbats). For MS-DOS users, this program has the
          additional advantage of being quite a bit faster than
          <b>Ghostscript</b>.</p>

          <p>Although <b>GoScript</b> can use any PostScript
          Type&#160;1 or Type&#160;3 font, the fonts supplied with
          <b>GoScript</b> are not in a standard PostScript format,
          so they cannot be used with other applications.</p>

          <p>On a 300dpi LaserJet printer, <b>GoScript</b> did not
          seem to align the baseline of bitmapped PostScript fonts
          with great accuracy. This is unfortunate, because the
          Computer Modern fonts created with MetaFont for TeX are
          bitmapped.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2897312"
              name="id2897312"></a>hp2xx</h3>
            </div>
          </div>

          <p><b>hp2xx</b><a id="id2897329" class="indexterm"
          name="id2897329"></a> converts scalable images in HPGL<a
          id="id2897338" class="indexterm" name="id2897338"></a>
          into other formats. HPGL is Hewlett-Packard's command
          language for controlling pen-based plotters. (HP's
          LaserJet series of printers, starting with the LaserJet
          III, also understand HPGL.)</p>

          <p>Many computer-aided drafting packages can save
          diagrams in HPGL format. Other software that works with
          scalable images is likely to be able to save in HPGL
          format as well. Because HPGL support is uncommon in
          printers, you will probably have to convert HPGL diagrams
          into another format to include them in your TeX
          documents. <b>hp2xx</b> output comes in three flavors:
          scalable graphics, TeX commands, and bitmapped
          graphics.</p>

          <div class="variablelist">
            <dl>
              <dt><span class="term">Scalable graphics</span></dt>

              <dd>
                <p>The supported scalable graphics formats are
                encapsulated PostScript and MetaFont source
                code.</p>
              </dd>

              <dt><span class="term">TeX commands</span></dt>

              <dd>
                <p>HPGL drawings composed of mostly straight lines
                can be rendered in TeX. <b>hp2xx</b>'s TeX output
                is designed for either the emTeX DVI drivers (it
                relies on \special commands to draw lines at
                arbitrary angles) or the <span
                class="emphasis"><em>epic</em></span> macros (which
                also use \special commands that must be supported
                by your DVI driver).</p>
              </dd>

              <dt><span class="term">Bitmapped graphics</span></dt>

              <dd>
                <p>Bitmaps can be produced in PBM<a id="id2897448"
                class="indexterm" name="id2897448"></a>, PCX<a
                id="id2897457" class="indexterm"
                name="id2897457"></a>, or PCL<a id="id2897465"
                class="indexterm" name="id2897465"></a> (HP
                LaserJet bitmap) formats. You can select image size
                and resolution to produce a bitmap that will print
                at the correct size for your printer.</p>
              </dd>
            </dl>
          </div>
        </div>
      </div>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a id="id2897483"
            name="id2897483"></a>Image Editors</h2>
          </div>
        </div>

        <p>Although many images<a id="id2897492" class="indexterm"
        name="id2897492"></a><a id="id2897500" class="indexterm"
        name="id2897500"></a> are available on the Net and
        commercially, it is very likely that you will want to
        create your own pictures and diagrams for some of the
        documents that you write. A few common drawing tools are
        described below. Table&#160;<a
        href="ch06.html#tab.drawingtools"
        title="Table&#160;6.2.&#160;Graphics Editing Packages">Table&#160;6.2</a>
        summarizes the packages discussed here. There are many more
        commercial packages for image editing---so many, in fact,
        that I'm not going to make any effort to describe them
        here.</p>

        <div class="table">
          <a id="tab.drawingtools" name="tab.drawingtools"></a>

          <p class="title"><b>Table&#160;6.2.&#160;Graphics Editing
          Packages</b></p>

          <table summary="Graphics Editing Packages" border="1">
            <colgroup>
              <col align="left" />
              <col align="center" />
              <col align="center" />
              <col align="left" />
            </colgroup>

            <thead>
              <tr>
                <th align="left">Program</th>

                <th align="center">Platforms</th>

                <th align="center">Cost</th>

                <th align="left">Purpose</th>
              </tr>
            </thead>

            <tbody>
              <tr>
                <td align="left">xfig</td>

                <td align="center">unix</td>

                <td align="center">free</td>

                <td align="left">Edit scalable Fig drawings</td>
              </tr>

              <tr>
                <td align="left">idraw</td>

                <td align="center">unix</td>

                <td align="center">free</td>

                <td align="left">Edit encapsulated PostScript
                images</td>
              </tr>

              <tr>
                <td align="left">tgif</td>

                <td align="center">unix</td>

                <td align="center">free</td>

                <td align="left">Edit scalable drawings</td>
              </tr>

              <tr>
                <td align="left">pixmap</td>

                <td align="center">unix</td>

                <td align="center">free</td>

                <td align="left">Edit X11 XBM and XPM files</td>
              </tr>

              <tr>
                <td align="left">texcad</td>

                <td align="center">MS-DOS</td>

                <td align="center">free</td>

                <td align="left">Edit LaTeX picture
                environments</td>
              </tr>

              <tr>
                <td align="left">xtexcad</td>

                <td align="center">unix</td>

                <td align="center">free</td>

                <td align="left">An X11 implementation of
                <b>texcad</b></td>
              </tr>
            </tbody>
          </table>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2897734"
              name="id2897734"></a>xfig</h3>
            </div>
          </div>

          <p><b>xfig</b><a id="id2897747" class="indexterm"
          name="id2897747"></a> is an editing tool for scalable
          images stored in the Fig format, described in the
          &#8220;<a href="ch06.html#sec.Fig"
          title="The epic and eepic styles">the section called
          &#8220;The epic and eepic styles&#8221;</a>&#8221;
          section of this chapter. An example of a drawing being
          constructed with <b>xfig</b> is shown in Figure&#160;<a
          href="ch06.html#fig.sd-xfig"
          title="Figure&#160;6.11.&#160;Editing an image with xfig">
          Figure&#160;6.11</a>.</p>

          <div class="figure">
            <a id="fig.sd-xfig" name="fig.sd-xfig"></a>

            <p class="title"><b>Figure&#160;6.11.&#160;Editing an
            image with xfig</b></p>

            <div class="mediaobject">
              <img src="tex.06.11.eps" />
            </div>
          </div>

          <p>The objects in Fig are arcs, circles, open and closed
          splines, ellipses, regular and irregular polygons,
          polylines, boxes, arc-boxes (boxes with rounded corners),
          text, and encapsulated PostScript. Compound objects can
          be constructed by binding these objects together.</p>

          <p>As a consequence of Fig's device independence, it can
          be translated into a number of different output formats.
          A completed Fig drawing can be rendered in PostScript,
          PiCTeX, \linebreak</p>

          <p>LaTeX's <tt>picture</tt> environment, MetaFont, or any
          of the pic, epic, eepic, box, eepicemu and \special
          environments. Figures can also be saved as X11 bitmaps.
          Not all of the output formats support all of the features
          of Fig. For example, arbitrarily sloped lines are not
          supported by the LaTeX <tt>picture</tt> environment, and
          encapsulated PostScript makes sense only for PostScript
          output.</p>

          <p><b>xfig</b> supports rotation, flipping, scaling, and
          duplication of objects. However, it imposes some
          limitations on the figures that you create. Objects that
          contain boxes, arc-boxes, circles, or ellipses can be
          rotated only in 90 degree increments. Text objects cannot
          be flipped over.</p>

          <p>When objects are placed on the &#8220;canvas,&#8221;
          you can elect to have <b>xfig</b> restrict their
          placement to $\frac{1}{16}$, $\frac{1}{4}$, or
          $\frac{1}{2}$-inch intervals. Restricted placement makes
          alignment of a large number of objects easier.
          Unrestricted placement is also allowed, and there is an
          alignment operator that can center (vertically and
          horizontally) any number of objects. A nonprinting grid
          is available in several sizes, independent of the
          restriction on object placement. In a similar way,
          <b>xfig</b> can be instructed to draw sloped lines within
          specific limits (for example, only allowing lines at
          slopes supported by LaTeX's <tt>picture</tt>
          environment). Precise numeric-coordinate placement of
          individual objects is also supported.</p>

          <p><b>xfig</b> provides access to the standard 35
          PostScript fonts as well as the standard LaTeX fonts.
          Eight-bit input, allowing access to international
          symbols, is also provided.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2897932"
              name="id2897932"></a>idraw</h3>
            </div>
          </div>

          <p><b>idraw</b><a id="id2897946" class="indexterm"
          name="id2897946"></a>, the InterViews drawing editor, is
          an X Windows application for editing encapsulated
          PostScript figures. Figure&#160;<a
          href="ch06.html#fig.sd-idraw"
          title="Figure&#160;6.12.&#160;Editing an image with idraw">
          Figure&#160;6.12</a> shows a drawing constructed with
          <b>idraw</b>.</p>

          <div class="figure">
            <a id="fig.sd-idraw" name="fig.sd-idraw"></a>

            <p class="title"><b>Figure&#160;6.12.&#160;Editing an
            image with idraw</b></p>

            <div class="mediaobject">
              <img src="FIXME:" />
            </div>
          </div>

          <p>The objects in <b>idraw</b> are lines, ellipses, open
          and closed splines, irregular polygons, polylines,
          rectangles, and text. Compound objects can be constructed
          by gluing these objects together. Fonts, brushes, and
          patterns can be customized and extended with X-defaults.
          <b>idraw</b> supports the full complement of rotation,
          flip, scale, and duplicate operations with or without
          grid lines and optional gravity.</p>

          <p>Unfortunately, the &#8220;gravity&#8221; option, which
          allows you to create horizontally and vertically aligned
          objects, is not preserved when a file is saved. It is
          very difficult to recover the alignment of a file when it
          has been saved and reloaded.</p>

          <p>Graphics in TIFF, encapsulated PostScript, X11 bitmap
          format, and Unidraw format can be imported into
          <b>idraw</b> figures.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2898057"
              name="id2898057"></a>tgif</h3>
            </div>
          </div>

          <p><b>tgif</b><a id="id2898071" class="indexterm"
          name="id2898071"></a>, like <b>xfig</b> and <b>idraw</b>,
          is an editor for scalable drawings. The
          &#8220;gif\,&#8221; in <b>tgif</b> has nothing to do with
          the bitmap GIF format. The captured screen in
          Figure&#160;<a href="ch06.html#fig.sd-tgif"
          title="Figure&#160;6.13.&#160;Editing an image with tgif">
          Figure&#160;6.13</a> shows a drawing being constructed
          with <b>tgif</b>.</p>

          <div class="figure">
            <a id="fig.sd-tgif" name="fig.sd-tgif"></a>

            <p class="title"><b>Figure&#160;6.13.&#160;Editing an
            image with tgif</b></p>

            <div class="mediaobject">
              <img src="FIXME:" />
            </div>
          </div>

          <p><b>tgif</b> stores objects as a set of Prolog
          &#8220;facts.&#8221; Several programs are provided for
          interpretation of <b>tgif</b> objects outside of
          <b>tgif</b>. Only four output types are built into
          <b>tgif</b>: PostScript<a id="id2898193"
          class="indexterm" name="id2898193"></a>, encapsulated
          PostScript (EPS)<a id="id2898202" class="indexterm"
          name="id2898202"></a>, X11 bitmaps<a id="id2898211"
          class="indexterm" name="id2898211"></a>, and X11
          pixmaps.</p>

          <p><b>tgif</b> has a number of interesting features
          including the ability to design hierarchical drawings in
          either a &#8220;top down&#8221; or &#8220;bottom
          up&#8221; manner, and the separation of an object's <span
          class="emphasis"><em>representation</em></span> from its
          <span class="emphasis"><em>instantiation</em></span>. In
          addition, arbitrary text-based attributes can be attached
          to each object. <b>tgif</b> uses these attributes in a
          number of ways, such as the execution of system commands
          based upon object attributes.</p>

          <p>The objects supported by <b>tgif</b> are arcs,
          ellipses, rectangles, rounded-corner rectangles,
          polylines, polygons, open and closed splines, text, X11
          bitmaps, some forms of X11 pixmaps, and encapsulated
          PostScript. These objects can be grouped together.</p>

          <p><b>tgif</b> supports only five fonts (Times, Courier,
          Helvetica, New Century Schoolbook, and Symbol) at a few
          fixed sizes. The fill patterns, line styles, text-styles,
          and other attributes are similarly fixed (although more
          generous in number).</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2898297"
              name="id2898297"></a>bitmap/pixmap</h3>
            </div>
          </div>

          <p><b>bitmap</b><a id="id2898311" class="indexterm"
          name="id2898311"></a> and <b>pixmap</b><a id="id2898325"
          class="indexterm" name="id2898325"></a> are standard
          tools distributed with the X11 Window System. The
          <b>pixmap</b> tool provides a superset of the functions
          in <b>bitmap</b>, including support for color images. An
          example of the <b>bitmap</b> program editing a small X
          icon is shown in Figure&#160;<a
          href="ch06.html#fig.xbitmap"
          title="Figure&#160;6.14.&#160;An example of bitmap editing an icon">
          Figure&#160;6.14</a>.</p>

          <div class="figure">
            <a id="fig.xbitmap" name="fig.xbitmap"></a>

            <p class="title"><b>Figure&#160;6.14.&#160;An example
            of bitmap editing an icon</b></p>

            <div class="mediaobject">
              <img src="FIXME:" />
            </div>
          </div>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2898403"
              name="id2898403"></a>Other Bitmap Editors</h3>
            </div>
          </div>

          <p>Free or inexpensive bitmap editors<a id="id2898413"
          class="indexterm" name="id2898413"></a> are available for
          almost every computer system. The selection of a
          particular package depends in large part on what kinds of
          diagrams you need to create.</p>

          <p>The primary disadvantage of bitmap editing is that it
          is difficult to create bitmaps large enough to be used at
          printer resolution. Even a &#8220;full screen&#8221;
          image on a high-resolution monitor is only a couple of
          inches across (at best) on a laser printer.</p>
        </div>

        <div class="section">
          <div class="titlepage">
            <div>
              <h3 class="title"><a id="id2898442"
              name="id2898442"></a>texcad/xtexcad</h3>
            </div>
          </div>

          <p><b>texcad</b><a id="id2898456" class="indexterm"
          name="id2898456"></a> is an MS-DOS program (distributed
          with emTeX) for editing LaTeX <tt>picture</tt>
          environments. <b>xtexcad</b><a id="id2898478"
          class="indexterm" name="id2898478"></a> is a similar
          program that runs under the X11 Window system.</p>

          <p>The LaTeX <tt>picture</tt> environment suffers from
          two limitations: it can only draw relatively simple
          diagrams, and it is very difficult to use. <b>texcad</b>
          and <b>xtexcad</b> remove the second limitation. With one
          of these programs, the LaTeX <tt>picture</tt> environment
          becomes a viable option for many diagrams.</p>
        </div>
      </div>

      <div class="section">
        <div class="titlepage">
          <div>
            <h2 class="title" style="clear: both"><a id="id2898527"
            name="id2898527"></a>Screen Dumps</h2>
          </div>
        </div>

        <p>If you are writing a document that describes a computer
        program, it is frequently desirable to include an image of
        the running program (a captured screen or screen dump<a
        id="id2898539" class="indexterm" name="id2898539"></a>) in
        the document. There are several ways that this can be
        accomplished<a id="id2898549" class="indexterm"
        name="id2898549"></a><a id="id2898556" class="indexterm"
        name="id2898556"></a><a id="id2898563" class="indexterm"
        name="id2898563"></a><a id="id2898570" class="indexterm"
        name="id2898570"></a><a id="id2898577" class="indexterm"
        name="id2898577"></a><a id="id2898584" class="indexterm"
        name="id2898584"></a><a id="id2898591" class="indexterm"
        name="id2898591"></a><a id="id2898598" class="indexterm"
        name="id2898598"></a>:</p>

        <div class="itemizedlist">
          <ul type="disc">
            <li>
              <p><b>Screen Thief</b> for MS-DOS</p>
            </li>

            <li>
              <p><b>GrabIt</b> for Microsoft Windows</p>
            </li>

            <li>
              <p><b>xv</b>, <b>Image Magick</b>, or <b>xwd</b> for
              X11 workstations</p>
            </li>

            <li>
              <p><b>PM-Cam</b>, <b>Nikon II</b>, or <b>pmjpeg</b>
              for OS/2</p>
            </li>
          </ul>
        </div>

        <p>The captured screen may be in a graphics format that is
        directly usable by your DVI driver or by <b>bm2font</b>. If
        not, one or more of the conversion programs described in
        &#8220;<a href="ch06.html#sec.picconv"
        title="Manipulating Images">the section called
        &#8220;Manipulating Images&#8221;</a>&#8221; earlier in
        this chapter will help you convert it into a usable
        format.</p>

        <p>The most common output devices for TeX are laser
        printers with a resolution of 300dpi or higher. By
        contrast, the resolution of a typical display is around
        80dpi. This discrepancy may require you to enlarge the
        bitmap image to make it legible in your document. For
        example, a 640x480 bitmap image is only about two inches
        wide when printed at 300dpi. In color images, <span
        class="emphasis"><em>dithering<a id="id2898740"
        class="indexterm" name="id2898740"></a></em></span>, which
        increases the number of dots used to represent each pixel
        in the original image, has a natural enlarging effect on
        the image, and this reduces the magnitude of the problem.
        However, some scaling may still be necessary.</p>

        <p>If the screen you want to include is only text (this
        includes IBM's line-drawing characters and other symbols),
        you can include it as a special &#8220;verbatim&#8221;
        environment if you have an appropriate font. The advantages
        of this method are that the actual screen text can be
        incorporated directly into your document (making the
        document more portable and easier to distribute, if that is
        a concern), and that the resulting document is small and
        easy to print. The disadvantage of this approach is that
        information about color is lost. If parts of the on-screen
        text appear in different colors for highlighting, using a
        pure-text approach may not produce acceptable results. For
        example, compare Figure&#160;<a
        href="ch06.html#fig.textshot"
        title="Figure&#160;6.15.&#160;A text mode screen dump.">Figure&#160;6.15</a>
        with the results in Figure&#160;<a
        href="ch02.html#fig.melatexcmp"
        title="Figure&#160;2.2.&#160;LaTeX as a compiler in Multi-Edit">
        Figure&#160;2.2</a> in Chapter&#160;<a href="ch02.html"
        title="Chapter&#160;2.&#160;Editing">Chapter&#160;2</a>,
        <span class="emphasis"><em><a href="ch02.html"
        title="Chapter&#160;2.&#160;Editing">Chapter&#160;2</a></em></span>.</p>

        <div class="figure">
          <a id="fig.textshot" name="fig.textshot"></a>

          <p class="title"><b>Figure&#160;6.15.&#160;A text mode
          screen dump.</b></p>

          <div class="literallayout">
            <p>\screendump{metext}</p>
          </div>
        </div>

        <p>To insert pure-text screen captures, you have to have a
        font with exactly the same encoding as the text on the
        screen (for more information about font encodings, see the
        section &#8220;<a href="ch05.html#sec.fonts.encodingvec"
        title="Declaring a family">the section called
        &#8220;Declaring a family&#8221;</a>&#8221; in
        Chapter&#160;<a href="ch05.html"
        title="Chapter&#160;5.&#160;Fonts">Chapter&#160;5</a>,
        <span class="emphasis"><em><a href="ch05.html"
        title="Chapter&#160;5.&#160;Fonts">Chapter&#160;5</a></em></span>).
        On MS-DOS and OS/2 PCs, this is sometimes a problem because
        the IBM line drawing characters do not appear in most
        fonts. Several commercial fonts provide the appropriate
        character set, but so do the freely available IBM Courier
        fonts (distributed by IBM for the X11 Consortium). The IBM
        Courier fonts are PostScript Type&#160;1 fonts. If you need
        TeX <tt>PK</tt> fonts, use the <b>ps2pk</b> program as
        described in the section called &#8220;<a
        href="ch05.html#sec.t1fonts" title="Math Fonts in TeX">the
        section called &#8220;Math Fonts in TeX&#8221;</a>&#8221;
        in Chapter&#160;<a href="ch05.html"
        title="Chapter&#160;5.&#160;Fonts">Chapter&#160;5</a>.</p>

        <p>To use the IBM Courier fonts to print captured screens,
        you will need <tt>AFM</tt> files<a id="id2898938"
        class="indexterm" name="id2898938"></a> that have the
        correct encoding vector. <tt>AFM</tt> files with the IBM
        OEM character set encoding are available from the CTAN
        archives in the directory <tt>fonts/courier</tt>.</p>

        <p>After the appropriate font is available, you are almost
        ready to reproduce a captured screen. One problem remains:
        TeX cannot \input binary files. Because screen dumps
        frequently contain characters from outside the normal ASCII
        range, they must be considered binary in this context. To
        overcome this difficulty, it is necessary to process the
        screen dump and convert it into a text file. The
        <b>Perl</b> script in Example&#160;<a
        href="apd.html#ex.txt2verb"
        title="Example&#160;D.8.&#160;txt2verb.pl">Example&#160;D.8</a>
        in Appendix&#160;<a href="apd.html"
        title="Appendix&#160;D.&#160;Long Examples">Appendix&#160;D</a>,
        <span class="emphasis"><em><a href="apd.html"
        title="Appendix&#160;D.&#160;Long Examples">Appendix&#160;D</a></em></span>,
        will perform this conversion.</p>

        <p>After conversion, the TeX code shown in Example&#160;<a
        href="ch06.html#ex.texscreencapt"
        title="Example&#160;6.9.&#160;Script for Inserting a Captured Text Screen">
        Example&#160;6.9</a> will insert the screen. This code will
        work in both Plain TeX and LaTeX. Similar code can be
        written for other formats. Figure&#160;<a
        href="ch06.html#fig.textshot"
        title="Figure&#160;6.15.&#160;A text mode screen dump.">Figure&#160;6.15</a>
        was produced with this code.</p>

        <div class="example">
          <a id="ex.texscreencapt" name="ex.texscreencapt"></a>

          <p class="title"><b>Example&#160;6.9.&#160;Script for
          Inserting a Captured Text Screen</b></p>
<pre class="programlisting">
<a type="simple" show="embed" actuate="onLoad"
href="screendump.sty"></a>
</pre>
        </div>
      </div>

      <div class="footnotes">
        <br />
        <hr width="100" align="left" />

        <div class="footnote">
          <p><sup>[<a id="ftn.id2884093" name="ftn.id2884093"
          href="#id2884093">79</a>]</sup> {Some devices, most
          notably plotters, don't work this way. They really draw
          lines and curves with physical pens.}</p>
        </div>

        <div class="footnote">
          <p><sup>[<a id="ftn.id2891727" name="ftn.id2891727"
          href="#id2891727">80</a>]</sup> And probably any other
          format derived from Plain TeX.</p>
        </div>

        <div class="footnote">
          <p><sup>[<a id="ftn.id2892937" name="ftn.id2892937"
          href="#id2892937">81</a>]</sup> {<b>dvimsp</b> has been
          replaced by <b>dvidot</b> in the beta test
          distributions.}</p>
        </div>

        <div class="footnote">
          <p><sup>[<a id="ftn.id2893346" name="ftn.id2893346"
          href="#id2893346">82</a>]</sup> {In fairness, the degree
          of difficulty depends on the platform you work on. Some
          systems, like the Amiga, have very good support for
          PostScript output, even on non-PostScript devices.
          Alternatively, if you have a fast computer, the extra
          translation through <b>Ghostscript</b> may be quite
          painless.}</p>
        </div>

        <div class="footnote">
          <p><sup>[<a id="ftn.id2893914" name="ftn.id2893914"
          href="#id2893914">83</a>]</sup> {The <tt>epsfig.sty</tt>
          style relies on <tt>epsf.tex</tt><a id="id2893934"
          class="indexterm" name="id2893934"></a> to provide
          low-level support for including EPS images. The
          <tt>epsf.tex</tt> file is part of the <b>dvips</b>
          distribution.}</p>
        </div>

        <div class="footnote">
          <p><sup>[<a id="ftn.id2896732" name="ftn.id2896732"
          href="#id2896732">84</a>]</sup> {Under MS-DOS and OS/2,
          remember that the command processor performs variable
          substitution with the percent sign. Use two consecutive
          percent signs.}</p>
        </div>
      </div>
    </div>

    <div class="navfooter">
      <table width="100%" summary="Navigation table">
        <tr>
          <td width="40%" align="left"><a
          title="Chapter&#160;5.&#160;Fonts" href="ch05.html"><img
          src="figures/nav-prev.png" alt="Prev" border="0" />
          </a>&#160;</td>

          <td width="20%" align="center"><a title="Making TeX Work"
          href="index.html"><img src="figures/nav-home.png"
          alt="Home" border="0" /></a></td>

          <td width="40%" align="right">&#160;<a
          title="Chapter&#160;7.&#160;International Considerations"
          href="ch07.html"><img src="figures/nav-next.png"
          alt="Next" border="0" /></a></td>
        </tr>

        <tr>
          <td width="40%" align="left">
          Chapter&#160;5.&#160;Fonts&#160;</td>

          <td width="20%" align="center"><a
          title="Part&#160;II.&#160;Elements of a Complex Document"
          href="pt02.html"><img src="figures/nav-up.png" alt="Up"
          border="0" /></a></td>

          <td width="40%" align="right">
          &#160;Chapter&#160;7.&#160;International
          Considerations</td>
        </tr>
      </table>
    </div>
  </body>
</html>