Remove more Java crap ...

Author: SergeStinckwich

DataMining.tex

@@ -906,17 +906,14 @@ \section{Covariance clusters}
 adapts itself to the shape of each cluster. As the algorithm
 progresses the metric changes dynamically.
 
-\subsection{Covariance clusters --- General implementation}
 \marginpar{Figure \ref{fig:dataminingclasses} with the boxes {\bf
 CovarianceCluster} grayed.} Covariance clusters need little
 implementation. All tasks are delegated to a Mahalanobis center
 described in section \ref{sec:mahalanobis}. Listing
-\ref{ls:mahacluster} shows the Smalltalk implementation and the
-Java implementation is shown in listing \ref{lj:mahacluster}.
+\ref{ls:mahacluster} shows the Smalltalk implementation.
 
 \begin{listing} Smalltalk covariance cluster \label{ls:mahacluster}
 \input{Smalltalk/DataMining/DhbCovarianceCluster}
 \end{listing}
 
-
 \ifx\wholebook\relax\else\end{document}\fi

Estimation.tex

@@ -594,15 +594,7 @@ \subsection{Weighted point implementation}
 the end of the name for Smalltalk --- implements the computation
 of one term of the sum in equation \ref{eq:defchitest}. The
 argument of the method is any object implementing the behavior of
-a one-variable function defined in section \ref{sec:function}. In
-Java one can use the same method name to define a similar method
-to compute the terms of the sum of equation
-\ref{eq:defchitestcmp}: in this case, the argument of the method
-is another weighted point. This is not possible in Smalltalk,
-which cannot distinguish the types of the arguments. Thus, for
-Smalltalk the second method must have a different name: {\tt
-chi2ComparisonContribution:}. Here Java marks a point over
-Smalltalk.
+a one-variable function defined in section \ref{sec:function}.
 
 Creating instances of the classes can be done in many ways. The
 fundamental method takes as arguments $x_i$, $y_i$ and the weight

FloatingPointSimulation.tex

@@ -6,8 +6,6 @@
 \begin{document}
 \fi
 
-
-
 \chapter{Decimal floating-point simulation}
 \label{ch-fpSimul} The class {\tt DhbDecimalFloatingNumber} is
 intended to demonstrate rounding problems with floating-point
@@ -21,11 +19,6 @@ \chapter{Decimal floating-point simulation}
 it is this model can be used to illustrate rounding problems to
 beginners. This class is only intended for didactical purposes.
 
-Only the Smalltalk implementation is given here, as Java does not
-lend itself to operator overloading. Moreover, fraction arithmetic
-is not available in Java. Thus, making an equivalent class would
-require much more code.
-
 Instances of the class are created with the method {\tt new:}
 supplying any number as argument. For example, $${\tt
 DhbDecimalFloatingNumber new: 3.141592653589793238462643}$$

LinearAlgebra.tex

@@ -960,7 +960,7 @@ \subsection{LUP decomposition --- General implementation}
 \end{description}
 
 The instance variable {\tt permutation} is set to undefined ({\tt
-nil} in Smalltlak, {\tt null} in Java) at initialization time by
+nil} in Smalltalk) at initialization time by
 default. It is used to check whether the decomposition has already
 been made or not.
 
@@ -1075,7 +1075,6 @@ \section{Computing the determinant of a matrix}
 the determinant is needed . The initial parity is 1. Each
 additional permutation of the rows multiplies the parity by -1.
 
-\subsection{Computing the determinant of matrix --- General implementation}
 Our implementation uses the fact that objects of the class {\tt
 Matrix} have an instance variable in which the LUP decomposition
 is kept. This variable is initialized using lazy initialization:
@@ -1090,7 +1089,6 @@ \subsection{Computing the determinant of matrix --- General implementation}
 product by the parity of the permutation to obtain the final
 result.
 
-\subsection{Computing the determinant of matrix implementation}
 Listing \ref{ls:determinant} shows the methods of classes {\tt
 DhbMatrix} and {\tt DhbLUPDecomposition} needed to compute a
 matrix determinant.
@@ -1100,14 +1098,6 @@ \subsection{Computing the determinant of matrix implementation}
 \input{Smalltalk/LinearAlgebra/DhbLUPDecomposition(DhbMatrixDeterminant)}
 \end{listing}
 
-
-\subsection{Computing the determinant of matrix --- Java implementation}
-The code computing the determinant of a matrix consists of the
-method {\tt determinant} of the class {\tt Matrix} (\cf listing
-\ref{ls:matrix}) and the method {\tt determinant} of the class
-{\tt LUPDecomposition} (\cf listing \ref{lj:lup}).
-
-
 \section{Matrix inversion}
 \label{sec:matrixinversion} The inverse of a square matrix ${\bf
 A}$ is denoted ${\bf A}^{-1}$. It is defined by the following

Series.tex

@@ -47,42 +47,20 @@ \section{Introduction}
 functions, which are very important to compute probabilities: the
 incomplete gamma function and the incomplete beta function.
 
-For illustrative purposes, the implementation in Smalltalk is
-using a different architecture from the one used by the Java
-implementation. It should be noted that each implementation could
-have been implemented in the other language. Figure
-\ref{fig:StSeriesClass} shows the class diagram of the Smalltalk
-implementation. Figure \ref{fig:JvSeriesClass} shows the class
-diagram of the Java implementation.
+Figure \ref{fig:StSeriesClass} shows the class diagram of the Smalltalk
+implementation.
 
 \begin{figure}
 \centering\includegraphics[width=11cm]{Figures/SeriesClassDiagram}
 \caption{Smalltalk class diagram for infinite series and continued
 fractions}\label{fig:StSeriesClass}
 \end{figure}
-\begin{figure}
-\centering\includegraphics[width=11cm]{Figures/SeriesClassDiagramJ}
-\caption{Java class diagram for infinite series and continued
-fractions}\label{fig:JvSeriesClass}
-\end{figure}
 
 The Smalltalk implementation uses two general-purpose classes to
 implement an infinite series and a continued fraction
 respectively. Each class then use a \patstyle{Strategy} pattern
 class \cite{GoF} to compute each term of the expansion.
 
-The Java implementation uses two abstract classes to implement an
-infinite series and a continued fraction respectively. Each
-concrete implementation necessitates the creation of a concrete
-subclass.
-
-In spite of the difference in architecture, the reader can verify
-on each class diagram that the number of classes needed for a
-concrete implementation is the same in each case.
-
-An interesting exercise for the reader is to implement the
-architecture presented in Java in Smalltalk and {\it vice versa}.
-
 \section{Infinite series}
 Many functions are defined with an infinite series, that is a sum
 of an infinite number of terms. The most well known example is the

Statistics.tex

@@ -1125,30 +1125,30 @@ \section{Probability distributions}
 used in this book. Other important distributions are presented in
 appendix \ref{ch:distributions}.
 
-\subsection{Probability distributions --- General implementation}
+\subsection{Probability distributions --- Smalltalk implementation}
 \marginpar{Figure \ref{fig:statisticsclasses} with the boxes {\bf
 ProbabilityDensity} and {\bf
 ProbabilityDensityWithUnknownDistribution} grayed.} Table
 \ref{tb:distrgenimpl} shows the description of the public methods
-of the implementations of both languages.
+of the implementation.
 \begin{table}[h]
   \centering
   \caption{Public methods for probability density functions}
   \label{tb:distrgenimpl}
 \vspace{1 ex}
-\begin{tabular}{|l | l | l|} \hline
-  Description & \hfil Smalltalk & \hfil Java \\ \hline
-  $P\left(x\right)$ & {\tt value:} & {\tt value(double)} \\
-  $F\left(x\right)$ & {\tt distributionValue:} & {\tt distributionValue(double)} \\
-  $F\left(x_1,x_2\right)$ & {\tt acceptanceBetween:and:} & {\tt distributionValue(double,double)} \\
-  $F^{-1}\left(x\right)$ & {\tt inverseDistributionValue:} & {\tt inverseDistributionValue(double)} \\
-  $x^{\dag}$ & {\tt random} & {\tt random()} \\
+\begin{tabular}{|l | l |} \hline
+  Description & \hfil Smalltalk \\ \hline
+  $P\left(x\right)$ & {\tt value:} \\
+  $F\left(x\right)$ & {\tt distributionValue:} \\
+  $F\left(x_1,x_2\right)$ & {\tt acceptanceBetween:and:} \\
+  $F^{-1}\left(x\right)$ & {\tt inverseDistributionValue:} \\
+  $x^{\dag}$ & {\tt random} \\
   \hline
-  $\bar{x}$ & {\tt average} & {\tt average()} \\
-  $\sigma^2$ & {\tt variance} & {\tt variance()} \\
-  $\sigma$ & {\tt standardDeviation} & {\tt standardDeviation()} \\
-  skewness & {\tt skewness} & {\tt skewness()} \\
-  kurtosis & {\tt kurtosis} & {\tt kurtosis()} \\
+  $\bar{x}$ & {\tt average} \\
+  $\sigma^2$ & {\tt variance} \\
+  $\sigma$ & {\tt standardDeviation} \\
+  skewness & {\tt skewness} \\
+  kurtosis & {\tt kurtosis} \\
   \hline
 \end{tabular}
 $\dag$ $x$ represents the random variable itself. In other words,
@@ -1174,7 +1174,6 @@ \subsection{Probability distributions --- General implementation}
 must be provided to create a function (as defined in section
 \ref{sec:function}) for the distribution function.
 
-\subsection{Probability distributions --- Smalltalk implementation}
 Listing \ref{ls:probdistr} shows the implementation of a general
 probability density distribution in Smalltalk. The class {\tt
 DhbProbabilityDensity} is an abstract implementation. Concrete