More complete change.

chapters/synchronous.tex

@@ -537,26 +537,27 @@ \section{Clocked State Variables}\label{clocked-state-variables}
 \hline
 \end{tabular}
 \end{center}
-\begin{nonnormative}
-At a clock tick only the (previous) values of the clocked state variables are needed to compute the new values of all clocked variables on that clock.
-This roughly corresponds to state variables in continuous time.
-
-In addition \lstinline!previous! can be applied to parameters returning their value.
-\end{nonnormative}
 
 \begin{operatordefinition}[previous]
 \begin{synopsis}\begin{lstlisting}
 previous($u$)
 \end{lstlisting}\end{synopsis}
 \begin{semantics}
-The input argument $u$ is a component expression (\cref{def:component-expression}), that variable is then a clocked state variable, or a parameter expression.
-The return argument has the same type as the input argument.
+The input argument $u$ is a component expression (\cref{def:component-expression}).
+If $u$ is a parameter its value is returned.
+
+Otherwise:
 Input and return arguments are on the same clock.
 At the first tick of the clock of $u$ or after a reset transition (see \cref{reset-handling}), the start value of $u$ is returned, see \cref{initialization-of-clocked-partitions}.
 At subsequent activations of the clock of $u$, the value of $u$ from the previous clock activation is returned.
 \end{semantics}
 \end{operatordefinition}
 
+\begin{nonnormative}
+At a clock tick only the (previous) values of the clocked state variables are needed to compute the new values of all clocked variables on that clock.
+This roughly corresponds to state variables in continuous time.
+\end{nonnormative}
+
 \section{Partitioning Operators}\label{partitioning-operators}
 
 A set of \emph{clock conversion operators} together act as boundaries