README-strip-blank.html

   1 <h5>Stripping blank elements</h5>
   2
   3 <p>Stripping blank elements has been implemented, i.e., elements that
   4   have no real text inside.
   5   This can enhance the readabilty of output considerably, especially
   6   when converting some automatically generated HTMLs.</p>
   7
   8 <p>By default, this only applies to tags for
   9   which this makes sense, i.e., FONT-like modifiers such as STRONG,
  10   EM, B, I, FONT. There are some other similar tags (such as CODE, TT,
  11   U, S) whose formatting can be important even on blanks and hence
  12   they are not stripped by default.</p>
  13
  14 <p>Note that, as a result, text formatting markup around images is stripped (which
  15   is good, because the markup is useless there).</p>
  16
  17 <h6>Optional stripping of blank elements</h6>
  18
  19 <p><code>--strip-blank-underline</code> command-line switch turns
  20   stripping of blank underlines (U tags) on, too. This can be useful for
  21   some generated HTMLs with meaningless underline.</p>
  22
  23 <h6>On implementation and future work</h6>
  24
  25 <p>The transformation is implemented on the level of HTML structure.
  26   I considered also implementing it on the level of the resulting
  27   wiki-text. On HTML level, it's straightforward that an image has no
  28   real text, and hence the formatting around it is to be stripped. On
  29   the level of plain text with wiki-markup, that wouldn't be so
  30   straightforward.</p>
  31
  32 <p>However, there is a related kind of pretty transformation that
  33   would be easier to perform on a level closer to the resulting
  34   wiki-text. Namely, "pushing" space from the edges out of the
  35   formatting. For example, something like <code>**Attention:&nbsp;**</code>
  36   should become <code>**Attention:** </code> (with the space being
  37   preserved outside). This can be easily implemented on the level of
  38   translation rules, we already have translated the content into
  39   wiki-text. Think of it as a variant of the
  40   <code>trim&nbsp;=&gt;&nbsp;both</code> rule, which would rather copy
  41   the space to the outside. This is for future work.
  42   </p>