Lout 3.17.

git-svn-id: http://svn.savannah.nongnu.org/svn/lout/trunk@2 9365b830-b601-4143-9ba8-b4a8e2c3339c

1 files changed, 61 insertions, 0 deletions

diff --git a/doc/expert/pre_bend b/doc/expert/pre_bend
new file mode 100644
index 0000000..8f742a9
--- /dev/null
+++ b/doc/expert/pre_bend
@@ -0,0 +1,61 @@
+@Section
+   @Tag { backend }
+   @Title { "@BackEnd" and the PlainText and PDF back ends }
+@Begin
+@PP
+backend.sym @Index { @@BackEnd symbol }
+The @@BackEnd symbol, which takes no parameters, has for its result a
+string naming the back end currently in use.  Three back ends are available,
+PostScript, PDF and PlainText.  The symbol is generally used like this:
+@ID @Code {
+"@BackEnd @Case {"
+"    PlainText @Yield { ... }"
+"    PostScript @Yield { ... }"
+"    PDF @Yield { ... }"
+"}"
+}
+to obtain different objects depending on the back end.  No @Code else
+is required since these are the only possible values.
+@PP
+When a @Code "@Case" symbol has @Code "@BackEnd" for its left parameter
+and the left parameter of each @Code "@Yield" symbol within it consists
+of a sequence of one or more literal words (including {@Code else}), Lout
+will optimize by evaluating the @Code "@Case" symbol at the time it is
+read.  This optimization ensures that there is only a small once-only
+performance penalty for multiple back ends, and it permits these
+@Code "@Case" symbols (but no other symbols) to appear within the object
+following @Code "@Include" and @Code "@PrependGraphic" symbols.
+@PP
+The PlainText back end differs from the PostScript one in two main
+respects.  First, there is effectively just one font:  although all
+the font commands work exactly as usual, they don't actually change
+anything.  Each character in this font is taken to be one tenth
+of one inch wide and 20 points high.  Second, the output is an
+ordinary text file, not a PostScript file.
+@PP
+Clearly, with ordinary text output the possibility of advanced graphics
+features such as rotation and scaling is curtailed.  Nevertheless, all
+symbols have well-defined (possibly null) effects in the PlainText
+back end, so there is no additional danger of crashing the system or
+obtaining grossly unreasonable output by a change to PlainText.
+@PP
+The PlainText back end is obtained by the @Code "-p" option to Basser
+Lout.  The character size can be changed by adding two lengths to
+the @Code "-p" option, like this:
+@ID @Code "lout -p0.1i12p ..."
+which invokes the PlainText back end with each character being 0.1
+inches wide and 12 points high.  However, experience suggests that
+the best approach is to define all horizontal lengths as multiples of
+the @Code "s" unit (the width of a space, hence the width of all
+characters) and to define all vertical lengths as multiples of the
+@Code "f" unit (the font size, equal to the height of every character),
+and not to change the character size in the command line.
+@PP
+There is a @Code "-P" option which is identical with the @Code "-p"
+option except that it inserts a form-feed character between each two
+components of the output, but not before the first or after the last.
+@PP
+The PDF back end is obtained by typing {@Code "lout -Z"}.  It is similar
+to PostScript but much more limited in functionality.  Consult a separate
+document distributed with Lout for further information.
+@End @Section