docs/usermanual-clusters.xml

<chapter id="clusters">
<sect1 id="clusters">
  <title>Clusters</title>
  <para>
    In shaping text, a <emphasis>cluster</emphasis> is a sequence of
    code points that needs to be treated as a single, indivisible unit.
  </para>
  <para>
    When you add text to a HB buffer, each character is associated with
    a <emphasis>cluster value</emphasis>. This is an arbitrary number as
    far as HB is concerned.
  </para>
  <para>
    Most clients will use UTF-8, UTF-16, or UTF-32 indices, but the
    actual number does not matter. Moreover, it is not required for the
    cluster values to be monotonically increasing, but pretty much all
    of HB's tests are performed on monotonically increasing cluster
    numbers. Nevertheless, there is no such assumption in the code
    itself. With that in mind, let's examine what happens with cluster
    values during shaping under each cluster-level.
  </para>
  <para>
    HarfBuzz provides three <emphasis>levels</emphasis> of clustering
    support. Level 0 is the default behavior and reproduces the behavior
    of the old HarfBuzz library. Level 1 tweaks this behavior slightly
    to produce better results, so level 1 clustering is recommended for
    code that is not required to implement backward compatibility with
    the old HarfBuzz.
  </para>
  <para>
    Level 2 differs significantly in how it treats cluster values.
    Levels 0 and 1 both process ligatures and glyph decomposition by
    merging clusters; level 2 does not.
  </para>
  <para>
    The conceptual model for what the cluster values mean, in levels 0
    and 1, is this:
  </para>
  <itemizedlist spacing="compact">
    <listitem>
      <para>
        the sequence of cluster values will always remain monotone
      </para>
    </listitem>
    <listitem>
      <para>
        each value represents a single cluster
      </para>
    </listitem>
    <listitem>
      <para>
        each cluster contains one or more glyphs and one or more
        characters
      </para>
    </listitem>
  </itemizedlist>
  <para>
    Assuming that initial cluster numbers were monotonically increasing
    and distinct, then all adjacent glyphs having the same cluster
    number belong to the same cluster, and all characters belong to the
    cluster that has the highest number not larger than their initial
    cluster number. This will become clearer with an example.
  </para>
</sect1>
<sect1 id="a-clustering-example-for-levels-0-and-1">
  <title>A clustering example for levels 0 and 1</title>
  <para>
    Let's say we start with the following character sequence and cluster
    values:
  </para>
  <programlisting>
   A,B,C,D,E
   0,1,2,3,4
</programlisting>
  <para>
    We then map the characters to glyphs. For simplicity, let's assume
    that each character maps to the corresponding, identical-looking
    glyph:
  </para>
  <programlisting>
   A,B,C,D,E
   0,1,2,3,4
</programlisting>
  <para>
    Now if, for example, <literal>B</literal> and <literal>C</literal>
    ligate, then the clusters to which they belong &quot;merge&quot;.
    This merged cluster takes for its cluster number the minimum of all
    the cluster numbers of the clusters that went in. In this case, we
    get:
  </para>
  <programlisting>
   A,BC,D,E
   0,1 ,3,4
</programlisting>
  <para>
    Now let's assume that the <literal>BC</literal> glyph decomposes
    into three components, and <literal>D</literal> also decomposes into
    two. The components each inherit the cluster value of their parent:
  </para>
  <programlisting>
   A,BC0,BC1,BC2,D0,D1,E
   0,1  ,1  ,1  ,3 ,3 ,4
</programlisting>
  <para>
    Now if <literal>BC2</literal> and <literal>D0</literal> ligate, then
    their clusters (numbers 1 and 3) merge into
    <literal>min(1,3) = 1</literal>:
  </para>
  <programlisting>
   A,BC0,BC1,BC2D0,D1,E
   0,1  ,1  ,1    ,1 ,4
</programlisting>
  <para>
    At this point, cluster 1 means: the character sequence
    <literal>BCD</literal> is represented by glyphs
    <literal>BC0,BC1,BC2D0,D1</literal> and cannot be broken down any
    further.
  </para>
</sect1>
<sect1 id="reordering-in-levels-0-and-1">
  <title>Reordering in levels 0 and 1</title>
  <para>
    Another common operation in the more complex shapers is when things
    reorder. In those cases, to maintain monotone clusters, HB merges
    the clusters of everything in the reordering sequence. For example,
    let's again start with the character sequence:
  </para>
  <programlisting>
   A,B,C,D,E
   0,1,2,3,4
</programlisting>
  <para>
    If <literal>D</literal> is reordered before <literal>B</literal>,
    then the <literal>B</literal>, <literal>C</literal>, and
    <literal>D</literal> clusters merge, and we get:
  </para>
  <programlisting>
   A,D,B,C,E
   0,1,1,1,4
</programlisting>
  <para>
    This is clearly not ideal, but it is the only sensible way to
    maintain monotone indices and retain the true relationship between
    glyphs and characters.
  </para>
</sect1>
<sect1 id="the-distinction-between-levels-0-and-1">
  <title>The distinction between levels 0 and 1</title>
  <para>
    So, the above is pretty much what cluster levels 0 and 1 do. The
    only difference between the two is this: in level 0, at the very
    beginning of the shaping process, we also merge clusters between
    base characters and all Unicode marks (combining or not) following
    them. E.g.:
  </para>
  <programlisting>
  A,acute,B
  0,1    ,2
</programlisting>
  <para>
    will become:
  </para>
  <programlisting>
  A,acute,B
  0,0    ,2
</programlisting>
  <para>
    This is the default behavior. We do it because Windows did it and
    old HarfBuzz did it, so this remained the default. But this behavior
    makes it impossible to color diacritic marks differently from their
    base characters. That's why in level 1 we do not perform this
    initial merging step.
  </para>
  <para>
    For clients, level 0 is more convenient if they rely on HarfBuzz
    clusters for cursor positioning. But that's wrong anyway: cursor
    positions should be determined based on Unicode grapheme boundaries,
    NOT shaping clusters. As such, level 1 clusters are preferred.
  </para>
  <para>
    One last note about levels 0 and 1. We currently don't allow a
    <literal>MultipleSubst</literal> lookup to replace a glyph with zero
    glyphs (i.e., to delete a glyph). But in some other situations,
    glyphs can be deleted. In those cases, if the glyph being deleted is
    the last glyph of its cluster, we make sure to merge the cluster
    with a neighboring cluster.
  </para>
  <para>
    This is, primarily, to make sure that the starting cluster of the
    text always has the cluster index pointing to the start of the text
    for the run; more than one client currently relies on this
    guarantee.
  </para>
  <para>
    Incidentally, Apple's CoreText does something else to maintain the
    same promise: it inserts a glyph with id 65535 at the beginning of
    the glyph string if the glyph corresponding to the first character
    in the run was deleted. HarfBuzz might do something similar in the
    future.
  </para>
</sect1>
<sect1 id="level-2">
  <title>Level 2</title>
  <para>
    Level 2 is a different beast from levels 0 and 1. It is simple to
    describe, but hard to make sense of. It simply doesn't do any
    cluster merging whatsoever. When things ligate or otherwise multiple
    glyphs turn into one, the cluster value of the first glyph is
    retained.
  </para>
  <para>
    Here are a few examples of why processing cluster values produced at
    this level might be tricky:
  </para>
  <sect2 id="ligatures-with-combining-marks">
    <title>Ligatures with combining marks</title>
    <para>
      Imagine capital letters are bases and lower case letters are
      combining marks. With an input sequence like this:
    </para>
    <programlisting>
  A,a,B,b,C,c
  0,1,2,3,4,5
</programlisting>
    <para>
      if <literal>A,B,C</literal> ligate, then here are the cluster
      values one would get under the various levels:
    </para>
    <para>
      level 0:
    </para>
    <programlisting>
  ABC,a,b,c
  0  ,0,0,0
</programlisting>
    <para>
      level 1:
    </para>
    <programlisting>
  ABC,a,b,c
  0  ,0,0,5
</programlisting>
    <para>
      level 2:
    </para>
    <programlisting>
  ABC,a,b,c
  0  ,1,3,5
</programlisting>
    <para>
      Making sense of the last example is the hardest for a client,
      because there is nothing in the cluster values to suggest that
      <literal>B</literal> and <literal>C</literal> ligated with
      <literal>A</literal>.
    </para>
  </sect2>
  <sect2 id="reordering">
    <title>Reordering</title>
    <para>
      Another tricky case is when things reorder. Under level 2:
    </para>
    <programlisting>
  A,B,C,D,E
  0,1,2,3,4
</programlisting>
    <para>
      Now imagine <literal>D</literal> moves before
      <literal>B</literal>:
    </para>
    <programlisting>
  A,D,B,C,E
  0,3,1,2,4
</programlisting>
    <para>
      Now, if <literal>D</literal> ligates with <literal>B</literal>, we
      get:
    </para>
    <programlisting>
  A,DB,C,E
  0,3 ,2,4
</programlisting>
    <para>
      In a different scenario, <literal>A</literal> and
      <literal>B</literal> could have ligated
      <emphasis>before</emphasis> <literal>D</literal> reordered; that
      would have resulted in:
    </para>
    <programlisting>
  AB,D,C,E
  0 ,3,2,4   
</programlisting>
    <para>
      There's no way to differentiate between these two scenarios based
      on the cluster numbers alone.
    </para>
    <para>
      Another problem happens with ligatures under level 2 if the
      direction of the text is forced to opposite of its natural
      direction (e.g. left-to-right Arabic). But that's too much of a
      corner case to worry about.
    </para>
  </sect2>
</sect1>
</chapter>