Clarify section separation in text format

While the old description ("The file should contain text data,
line-based, with single empty lines separating sections. The
lines themselves are column-based, with the pipe symbol (|) acting
as separator.") is precise and unambiguous, we did have some
misunderstandings in the past. So add the format of the empty
section (which makes sense to describe a cluster without any
instances) as a clarifying example.

Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Petr Pudlak <pudlak@google.com>

man/htools.rst

@@ -154,7 +154,11 @@ support all options. Some common options are:
   page **htools**\(1).
 
   The file should contain text data, line-based, with single empty lines
-  separating sections. The lines themselves are column-based, with the
+  separating sections. In particular, an empty section is described by
+  the empty string followed by the separating empty line, thus yielding
+  two consecutive empty lines. So the number of empty lines does matter
+  and cannot be changed arbitrarily.
+  The lines themselves are column-based, with the
   pipe symbol (``|``) acting as separator.
 
   The first section contains group data, with the following columns: