\PassOptionsToPackage{unicode=true}{hyperref} % options for packages loaded elsewhere \PassOptionsToPackage{hyphens}{url} % \documentclass[]{article} \usepackage{lmodern} \usepackage{amssymb,amsmath} \usepackage{ifxetex,ifluatex} \usepackage{fixltx2e} % provides \textsubscript \ifnum 0\ifxetex 1\fi\ifluatex 1\fi=0 % if pdftex \usepackage[T1]{fontenc} \usepackage[utf8]{inputenc} \usepackage{textcomp} % provides euro and other symbols \else % if luatex or xelatex \usepackage{unicode-math} \defaultfontfeatures{Ligatures=TeX,Scale=MatchLowercase} \fi % use upquote if available, for straight quotes in verbatim environments \IfFileExists{upquote.sty}{\usepackage{upquote}}{} % use microtype if available \IfFileExists{microtype.sty}{% \usepackage[]{microtype} \UseMicrotypeSet[protrusion]{basicmath} % disable protrusion for tt fonts }{} \IfFileExists{parskip.sty}{% \usepackage{parskip} }{% else \setlength{\parindent}{0pt} \setlength{\parskip}{6pt plus 2pt minus 1pt} } \usepackage{hyperref} \hypersetup{ pdftitle={CSci 658: Software Language Engineering Notes on Scala for Java Programmers}, pdfauthor={H. Conrad Cunningham}, pdfborder={0 0 0}, breaklinks=true} \urlstyle{same} % don't use monospace font for urls \usepackage{color} \usepackage{fancyvrb} \newcommand{\VerbBar}{|} \newcommand{\VERB}{\Verb[commandchars=\\\{\}]} \DefineVerbatimEnvironment{Highlighting}{Verbatim}{commandchars=\\\{\}} % Add ',fontsize=\small' for more characters per line \newenvironment{Shaded}{}{} \newcommand{\AlertTok}[1]{\textcolor[rgb]{1.00,0.00,0.00}{\textbf{#1}}} \newcommand{\AnnotationTok}[1]{\textcolor[rgb]{0.38,0.63,0.69}{\textbf{\textit{#1}}}} \newcommand{\AttributeTok}[1]{\textcolor[rgb]{0.49,0.56,0.16}{#1}} \newcommand{\BaseNTok}[1]{\textcolor[rgb]{0.25,0.63,0.44}{#1}} \newcommand{\BuiltInTok}[1]{#1} \newcommand{\CharTok}[1]{\textcolor[rgb]{0.25,0.44,0.63}{#1}} \newcommand{\CommentTok}[1]{\textcolor[rgb]{0.38,0.63,0.69}{\textit{#1}}} \newcommand{\CommentVarTok}[1]{\textcolor[rgb]{0.38,0.63,0.69}{\textbf{\textit{#1}}}} \newcommand{\ConstantTok}[1]{\textcolor[rgb]{0.53,0.00,0.00}{#1}} \newcommand{\ControlFlowTok}[1]{\textcolor[rgb]{0.00,0.44,0.13}{\textbf{#1}}} \newcommand{\DataTypeTok}[1]{\textcolor[rgb]{0.56,0.13,0.00}{#1}} \newcommand{\DecValTok}[1]{\textcolor[rgb]{0.25,0.63,0.44}{#1}} \newcommand{\DocumentationTok}[1]{\textcolor[rgb]{0.73,0.13,0.13}{\textit{#1}}} \newcommand{\ErrorTok}[1]{\textcolor[rgb]{1.00,0.00,0.00}{\textbf{#1}}} \newcommand{\ExtensionTok}[1]{#1} \newcommand{\FloatTok}[1]{\textcolor[rgb]{0.25,0.63,0.44}{#1}} \newcommand{\FunctionTok}[1]{\textcolor[rgb]{0.02,0.16,0.49}{#1}} \newcommand{\ImportTok}[1]{#1} \newcommand{\InformationTok}[1]{\textcolor[rgb]{0.38,0.63,0.69}{\textbf{\textit{#1}}}} \newcommand{\KeywordTok}[1]{\textcolor[rgb]{0.00,0.44,0.13}{\textbf{#1}}} \newcommand{\NormalTok}[1]{#1} \newcommand{\OperatorTok}[1]{\textcolor[rgb]{0.40,0.40,0.40}{#1}} \newcommand{\OtherTok}[1]{\textcolor[rgb]{0.00,0.44,0.13}{#1}} \newcommand{\PreprocessorTok}[1]{\textcolor[rgb]{0.74,0.48,0.00}{#1}} \newcommand{\RegionMarkerTok}[1]{#1} \newcommand{\SpecialCharTok}[1]{\textcolor[rgb]{0.25,0.44,0.63}{#1}} \newcommand{\SpecialStringTok}[1]{\textcolor[rgb]{0.73,0.40,0.53}{#1}} \newcommand{\StringTok}[1]{\textcolor[rgb]{0.25,0.44,0.63}{#1}} \newcommand{\VariableTok}[1]{\textcolor[rgb]{0.10,0.09,0.49}{#1}} \newcommand{\VerbatimStringTok}[1]{\textcolor[rgb]{0.25,0.44,0.63}{#1}} \newcommand{\WarningTok}[1]{\textcolor[rgb]{0.38,0.63,0.69}{\textbf{\textit{#1}}}} \setlength{\emergencystretch}{3em} % prevent overfull lines \providecommand{\tightlist}{% \setlength{\itemsep}{0pt}\setlength{\parskip}{0pt}} \setcounter{secnumdepth}{0} % Redefines (sub)paragraphs to behave more like sections \ifx\paragraph\undefined\else \let\oldparagraph\paragraph \renewcommand{\paragraph}[1]{\oldparagraph{#1}\mbox{}} \fi \ifx\subparagraph\undefined\else \let\oldsubparagraph\subparagraph \renewcommand{\subparagraph}[1]{\oldsubparagraph{#1}\mbox{}} \fi % set default figure placement to htbp \makeatletter \def\fps@figure{htbp} \makeatother \usepackage{caption} \DeclareCaptionLabelFormat{nolabel}{} \captionsetup{labelformat=nolabel} \title{CSci 658: Software Language Engineering\\ Notes on Scala for Java Programmers} \author{\textbf{H. Conrad Cunningham}} \date{\textbf{17 February 2018}} \begin{document} \maketitle { \setcounter{tocdepth}{4} \tableofcontents } \href{http://www.cs.olemiss.edu/~hcc}{H. Conrad Cunningham}\\ Professor of \href{https://www.cs.olemiss.edu}{Computer and Information Science}\\ \href{http://www.olemiss.edu}{University of Mississippi}\\ 211 Weir Hall\\ P.O. Box 1848\\ University, MS 38677\\ (662) 915-5358 \textbf{Advisory}: The HTML version of this document may require use of a browser that supports the display of MathML. A good choice as of Februlary 2018 is a recent version of Firefox from Mozilla. \hypertarget{notes-on-scala-for-java-programmers}{% \section{Notes on Scala for Java Programmers}\label{notes-on-scala-for-java-programmers}} \hypertarget{introduction}{% \subsection{Introduction}\label{introduction}} This is an introduction to Scala for programmers who have completed the equivalent of the Java-based Computer Science I-II-III (CSci 111-112-211) sequence at the University of Mississippi. I adapt and expand the document \href{http://docs.scala-lang.org/tutorials/scala-for-java-programmers.html}{``A Scala Tutorial for Java Programmers''} by Michel Schinz and Phillipp Haller to better meet the needs of my Scala-based courses (to explain some items more thoroughly, improve the narrative, and make the document accessible). \hypertarget{a-first-example-hello-world}{% \subsection{A First Example: Hello World}\label{a-first-example-hello-world}} A ``Hello, world!'' program is the obligatory first example to give when introducing a new language. We can write a program \texttt{HelloWorld} as follows in Scala: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{object}\NormalTok{ HelloWorld \{} \KeywordTok{def} \FunctionTok{main}\NormalTok{(args: Array[String]) \{} \FunctionTok{println}\NormalTok{(}\StringTok{"Hello, world!"}\NormalTok{)} \NormalTok{ \}} \NormalTok{ \}} \end{Highlighting} \end{Shaded} What Scala features do we use here in relation to Java? \begin{itemize} \item Keyword \texttt{object} declares a \emph{singleton object} named \texttt{HelloWorld}. An object is essentially a class with a single instance. The body of the object is enclosed in braces following the name. \item The keyword \texttt{def} introduces a method definition. \item In a declaration, a colon (\texttt{:}) separates the name from its type. \item Method \texttt{main} takes the command line arguments as its parameter, which is an array of strings. \item The \texttt{main} method is a procedure and, hence, has no return type declared. The body of the method is enclosed in braces following the method header. \item The \texttt{main} method is not declared as \texttt{static} as in Java. Static members do not exist in Scala. We can use singleton objects instead. \item The body of \texttt{main} has a single call to predefined method \texttt{println}. \end{itemize} \hypertarget{compiling-the-example}{% \subsubsection{Compiling the example}\label{compiling-the-example}} We can use the \texttt{scalac} command (similar to the \texttt{javac} command) to invoke the Scala compiler. If the above Scala program is stored in file \texttt{HelloWorld.scala}, we can compile it from the command line as follows: \begin{Shaded} \begin{Highlighting}[] \ExtensionTok{scalac}\NormalTok{ HelloWorld.scala} \end{Highlighting} \end{Shaded} The above compiles the Scala source file and generates a few class files in the current directory. File \texttt{HelloWorld.class} contains a class that can be executed. \hypertarget{running-the-example}{% \subsubsection{Running the example}\label{running-the-example}} We can use the \texttt{scala} command (similar to the \texttt{java} command) to execute the \texttt{main} method. Execution of the program prints the ``Hello, World'' string to the console. \begin{Shaded} \begin{Highlighting}[] \OperatorTok{>} \ExtensionTok{scala}\NormalTok{ -classpath . HelloWorld} \ExtensionTok{Hello}\NormalTok{, world!} \end{Highlighting} \end{Shaded} \hypertarget{interaction-with-java}{% \subsection{Interaction with Java}\label{interaction-with-java}} Scala code can interact with Java code. Package \texttt{java.lang} is imported by default and other packages can be imported explicitly. Consider a program to obtain and format the current date according to the conventions used in a specific country, say France. \begin{Shaded} \begin{Highlighting}[] \KeywordTok{import}\NormalTok{ java.}\FunctionTok{util}\NormalTok{.\{Date, Locale\}} \KeywordTok{import}\NormalTok{ java.}\FunctionTok{text}\NormalTok{.}\FunctionTok{DateFormat} \KeywordTok{import}\NormalTok{ java.}\FunctionTok{text}\NormalTok{.}\FunctionTok{DateFormat}\NormalTok{._} \KeywordTok{object}\NormalTok{ FrenchDate \{} \KeywordTok{def} \FunctionTok{main}\NormalTok{(args: Array[String]) \{} \KeywordTok{val}\NormalTok{ now = }\KeywordTok{new}\NormalTok{ Date} \KeywordTok{val}\NormalTok{ df = }\FunctionTok{getDateInstance}\NormalTok{(LONG, Locale.}\FunctionTok{FRANCE}\NormalTok{)} \FunctionTok{println}\NormalTok{(df format now)} \NormalTok{ \}} \NormalTok{ \}} \end{Highlighting} \end{Shaded} Java's class libraries such as \texttt{Date} and \texttt{DateFormat} can be called directly from Scala code. The Scala \texttt{import} statement is more powerful than Java's. It can: \begin{itemize} \item Import multiple classes by enclosing names in braces. The first line above imports the \texttt{Data} and \texttt{Locale} classes from Java package \texttt{java.util}. \item Import everything in a package or class by using an underscore character (\texttt{\_}) instead of a Java's asterisk (\texttt{*}). The third line imports all members of the \texttt{DateFormat} class, making static method \texttt{getDateInstance} and static field \texttt{LONG} visible. \end{itemize} The \texttt{main} method in the \texttt{FrenchDate} object: \begin{itemize} \item Creates an instance of Java's \texttt{Date} class, thus getting the current date \item Uses the imported static method \texttt{getDateInstance} to format the date appropriately for France using \texttt{Locale.France} \item Prints the current date formatted according to the localized \texttt{DateFormat} instance \end{itemize} The \texttt{main} method declares two local ``variables'' in its body: \texttt{now} and \texttt{df}. Note that: \begin{itemize} \item Both are declared with \texttt{val}. After initialization, Scala does not allow further assignments to a \texttt{val}. The alternative is \texttt{var}. Scala does allow assignments to a \texttt{var}; it is like an ordinary variable in Java. Although we cannot change the binding of a \texttt{val} to an object, Scala does allow the internal state of the object to be changed. \item Although Scala is statically typed like Java, neither of these variables are given explicit types. The types of the variables are {[}\emph{inferred} {]}\url{https://en.wikipedia.org/wiki/Type_inference}) by the compiler from the type of the initializing expression. \end{itemize} Scala methods taking one explicit argument can be written in infix syntax such as \begin{Shaded} \begin{Highlighting}[] \NormalTok{ df format now} \end{Highlighting} \end{Shaded} which is the same as the method call \begin{Shaded} \begin{Highlighting}[] \NormalTok{ df}\FunctionTok{.format}\NormalTok{(now)} \end{Highlighting} \end{Shaded} This feature has important consequences, as we discuss below. It is also possible to inherit from Java classes and implement Java interfaces directly in Scala. \hypertarget{everything-is-an-object}{% \subsection{Everything is an Object}\label{everything-is-an-object}} Scala is a pure object-oriented language. \emph{Everything} is an object, including numbers and functions. Scala does not distinguish primitive types (e.g., \texttt{boolean} and \texttt{int}) from reference types (e.g., objects). It enables us to manipulate functions as first-class values. \hypertarget{numbers-are-objects}{% \subsubsection{Numbers are objects}\label{numbers-are-objects}} Since numbers are objects, they also have methods. For example, the expression \begin{Shaded} \begin{Highlighting}[] \DecValTok{1}\NormalTok{ + }\DecValTok{2}\NormalTok{ * }\DecValTok{3}\NormalTok{ / x} \end{Highlighting} \end{Shaded} is equivalent to the expression \begin{Shaded} \begin{Highlighting}[] \NormalTok{ (}\DecValTok{1}\NormalTok{).+(((}\DecValTok{2}\NormalTok{).*(}\DecValTok{3}\NormalTok{))./(x))} \end{Highlighting} \end{Shaded} which shows the method calls explicitly. In Scala, the ``operator'' symbols (e.g., \texttt{+}, \texttt{*}, \texttt{/}) are valid identifiers. The parentheses around numbers are necessary because Scala's lexical analyzer uses a longest-match rule for tokens, breaking expression \begin{Shaded} \begin{Highlighting}[] \FloatTok{1.}\NormalTok{+(}\DecValTok{2}\NormalTok{)} \end{Highlighting} \end{Shaded} into the tokens \texttt{1.}, \texttt{+}, and \texttt{2}. This results in \texttt{1.} being interpreted as a \texttt{Double}. \hypertarget{functions-are-objects}{% \subsubsection{Functions are objects}\label{functions-are-objects}} Because Scala functions are objects, we can: \begin{itemize} \item pass a function as an argument to or return a function as a result from another function. That is, Scala has \href{https://en.wikipedia.org/wiki/Higher-order_function}{\emph{higher-order}} functions. \item store a function in a variable or data structure. That is, Scala's functions are \href{https://en.wikipedia.org/wiki/First-class_function}{\emph{first-class}} values. \end{itemize} These are key features of \href{https://en.wikipedia.org/wiki/Functional_programming}{\emph{functional programming}}. Consider the \texttt{Timer} program below. It includes a timer function named \texttt{oncePerSecond} that performs some action every second. The specific action performed is encoded as a \emph{call-back} function passed into \texttt{oncePerSecond} as an argument. \begin{Shaded} \begin{Highlighting}[] \KeywordTok{object}\NormalTok{ Timer \{} \KeywordTok{def} \FunctionTok{oncePerSecond}\NormalTok{(callback: () => Unit) \{} \KeywordTok{while}\NormalTok{ (}\KeywordTok{true}\NormalTok{) \{ }\FunctionTok{callback}\NormalTok{(); Thread sleep }\DecValTok{1000}\NormalTok{ \}} \NormalTok{ \}} \KeywordTok{def} \FunctionTok{timeFlies}\NormalTok{() \{} \FunctionTok{println}\NormalTok{(}\StringTok{"time flies like an arrow..."}\NormalTok{)} \NormalTok{ \}} \KeywordTok{def} \FunctionTok{main}\NormalTok{(args: Array[String]) \{} \FunctionTok{oncePerSecond}\NormalTok{(timeFlies)} \NormalTok{ \}} \NormalTok{ \}} \end{Highlighting} \end{Shaded} The type of the call-back function is \texttt{()\ =\textgreater{}\ Unit}. This means it is a function that takes no arguments and returns nothing. The type \texttt{Unit} is similar to \texttt{void} in C/C++. Function \texttt{oncePerSecond} calls the \texttt{Thread.sleep} method (from \texttt{java.lang}) and uses an infinite \texttt{while} loop to repeat the call-back action every second. The \texttt{main} function calls this timer function, passing a call-back function that prints the string ``time flies like an arrow\ldots{}'' on the console. The program endlessly prints this string every second. The program uses the predefined Scala method \texttt{println} instead the like-named Java method from \texttt{System.out}. \hypertarget{anonymous-functions}{% \subsubsection{Anonymous functions}\label{anonymous-functions}} The \texttt{Timer} program in the subsection above can be refined by replacing the function \texttt{timeFlies} by an \href{https://en.wikipedia.org/wiki/Anonymous_function}{\emph{anonymous function}} as shown below: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{object}\NormalTok{ TimerAnonymous \{} \KeywordTok{def} \FunctionTok{oncePerSecond}\NormalTok{(callback: () => Unit) \{} \KeywordTok{while}\NormalTok{ (}\KeywordTok{true}\NormalTok{) \{ }\FunctionTok{callback}\NormalTok{(); Thread sleep }\DecValTok{1000}\NormalTok{ \}} \NormalTok{ \}} \KeywordTok{def} \FunctionTok{main}\NormalTok{(args: Array[String]) \{} \FunctionTok{oncePerSecond}\NormalTok{(() =>} \FunctionTok{println}\NormalTok{(}\StringTok{"time flies like an arrow..."}\NormalTok{))} \NormalTok{ \}} \NormalTok{ \}} \end{Highlighting} \end{Shaded} In the anonymous function, a right arrow \texttt{=\textgreater{}} separates the function's argument list from its body expression. In this example, the argument list is empty, shown by the empty pair of parentheses on the left. \hypertarget{classes}{% \subsection{Classes}\label{classes}} Like Java, Scala is a class-based language. However, Scala classes can have parameters. Consider the following class \texttt{Complex} for representing complex numbers: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{class} \FunctionTok{Complex}\NormalTok{(real: Double, imaginary: Double) \{} \KeywordTok{def} \FunctionTok{re}\NormalTok{() = real} \KeywordTok{def} \FunctionTok{im}\NormalTok{() = imaginary} \NormalTok{ \}} \end{Highlighting} \end{Shaded} The \texttt{Complex} class takes two arguments, the real and imaginary parts of the complex number. In the \texttt{Complex} class, these become \texttt{val} fields of the object (i.e., class instance) that are only visible and accessible from inside the object. (This visibility is \texttt{private{[}this{]}}, a type of visibility that does not occur in Java.) The default constructor is built into the \texttt{class} syntax. We pass values for these arguments when we create an instance of class \texttt{Complex}, as follows: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{new} \FunctionTok{Complex}\NormalTok{(}\FloatTok{1.5}\NormalTok{, }\FloatTok{2.3}\NormalTok{)} \end{Highlighting} \end{Shaded} This causes all statements in the class definition to be executed. The \texttt{Complex} class defines function methods \texttt{re} and \texttt{im}, which provide access to the two parts of the complex number. That is, these are \emph{accessor}, or \emph{getter}, methods. We declare a function with an \texttt{=} between the function's header (i.e., name and parameter list) and its body. The body is an expression that is evaluated when the function is called. If the body itself consists of a sequence of expressions, we enclose it in braces. The value of last expression executed is the value of the body. In \texttt{Complex}, the compiler infers the return types for methods \texttt{re()} and \texttt{im()} by examining the right-hand sides and deducing that both return a value of type \texttt{Double}. The compiler gives an error message when it cannot infer the type of a method or variable. Some Scala programmers suggest that the types be omitted whenever possible and only inserted when necessary. However, your instructor considers it better software engineering practice to explicitly specify the types of all \emph{public} features of a class or package such as the \texttt{re()} and \texttt{im()} methods. But, for internal features of a class or method, it is convenient and safe to use type inference. For example, we did not give explicit types for the \texttt{now} and \texttt{df} values in the \texttt{FrenchDate} object shown in a previous section. \hypertarget{methods-without-arguments}{% \subsubsection{Methods without arguments}\label{methods-without-arguments}} To call the methods \texttt{re} and \texttt{im}, we must put an empty pair of parentheses after their names: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{object}\NormalTok{ ComplexNumbers \{} \KeywordTok{def} \FunctionTok{main}\NormalTok{(args: Array[String]) \{} \KeywordTok{val}\NormalTok{ c = }\KeywordTok{new} \FunctionTok{Complex}\NormalTok{(}\FloatTok{1.2}\NormalTok{, }\FloatTok{3.4}\NormalTok{)} \FunctionTok{println}\NormalTok{(}\StringTok{"imaginary part: "}\NormalTok{ + c.}\FunctionTok{im}\NormalTok{())} \NormalTok{ \}} \NormalTok{ \}} \end{Highlighting} \end{Shaded} We can eliminate the empty parameter list by defining \texttt{re} and \texttt{im} as methods \emph{without arguments} instead of \emph{with zero arguments}. We simply omit the parentheses in the method definition, as shown below: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{class} \FunctionTok{Complex}\NormalTok{(real: Double, imaginary: Double) \{} \KeywordTok{def}\NormalTok{ re = real} \KeywordTok{def}\NormalTok{ im = imaginary} \NormalTok{ \}} \end{Highlighting} \end{Shaded} \hypertarget{inheritance-and-overriding}{% \subsubsection{Inheritance and overriding}\label{inheritance-and-overriding}} All classes in Scala inherit from a superclass. If no super class is given explicitly, then the superclass is \texttt{scala.AnyRef} by default. Like Java, a Scala class inherits all methods from its superclass by default. Like Java, a Scala class may override individual methods by giving new definitions. Unlike Java, a Scala method that overrides a superclass method must be explicitly declared with the \texttt{override} modifier. This is to reduce the possibility of class overriding a superclass method by accident. For example, the\texttt{Complex} class can redefine the \texttt{toString} method inherited from \texttt{AnyRef}. \begin{Shaded} \begin{Highlighting}[] \KeywordTok{class} \FunctionTok{Complex}\NormalTok{(real: Double, imaginary: Double) \{} \KeywordTok{def}\NormalTok{ re = real} \KeywordTok{def}\NormalTok{ im = imaginary} \KeywordTok{override} \KeywordTok{def} \FunctionTok{toString}\NormalTok{() =} \StringTok{""}\NormalTok{ + re + (}\KeywordTok{if}\NormalTok{ (im < }\DecValTok{0}\NormalTok{) }\StringTok{""} \KeywordTok{else} \StringTok{"+"}\NormalTok{) + im + }\StringTok{"i"} \NormalTok{ \}} \end{Highlighting} \end{Shaded} The new definition of \texttt{toString()} overrides the definition for \texttt{Complex} objects. However, in doing so, it uses the default definitions of \texttt{toString()} for \texttt{Double} by its references to \texttt{re} and \texttt{im} in the string concatenation. \hypertarget{case-classes-and-pattern-matching}{% \subsection{Case Classes and Pattern Matching}\label{case-classes-and-pattern-matching}} In programming, we often use trees and other hierarchical data structures. We can illustrate how to implement a tree in Scala using a small calculator program for simple arithmetic expressions composed of sums, integer constants, and variables. Examples of such expressions are \texttt{1+2} and \texttt{(x+x)+(7+y)}. We can represent expressions naturally with a tree, where nodes are operations (e.g., addition) and leaves are values (e.g., constants or variables). In Java, we can represent such a tree using an abstract superclass for the trees and a concrete subclass for each kind of internal and leaf node. The abstract class defines the common interface to the tree and the concrete subclasses define the specific features of a node. In functional programming languages such as Haskell, we can define an \href{https://en.wikipedia.org/wiki/Algebraic_data_type}{\emph{algebraic data type}} for the same purpose. These types often enable us to express programs concisely by using pattern matching constructs. Scala combines the concepts of classes and algebraic data types into its \emph{case classes}. Consider the example: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{abstract} \KeywordTok{class}\NormalTok{ Tree } \KeywordTok{case} \KeywordTok{class} \FunctionTok{Sum}\NormalTok{(l: Tree, r: Tree) }\KeywordTok{extends}\NormalTok{ Tree } \KeywordTok{case} \KeywordTok{class} \FunctionTok{Var}\NormalTok{(n: String) }\KeywordTok{extends}\NormalTok{ Tree } \KeywordTok{case} \KeywordTok{class} \FunctionTok{Const}\NormalTok{(v: Int) }\KeywordTok{extends}\NormalTok{ Tree } \end{Highlighting} \end{Shaded} Here, we define \texttt{Sum}, \texttt{Var} and \texttt{Const} as case classes. These differ from from standard classes in several ways: \begin{itemize} \item We can create an instance without using the keyword \texttt{new} keyword, for example, we can write \texttt{Const(5)} instead of \texttt{new\ Const(5)}. \item The compiler automatically generates getter functions for the constructor parameters. That is, it is possible to get the value of the \texttt{v} constructor parameter of some instance \texttt{c} of class \texttt{Const} just by writing \texttt{c.v}. \item The compiler supplies default definitions for the methods \texttt{equals} and \texttt{hashCode}. These work on the \emph{structure} of the instances and not on their identities. \item The compiler provides a default definition for method \texttt{toString}, which prints the value in a ``source'' form. For example, the tree for expression \texttt{x+1} prints as \texttt{Sum(Var(x),Const(1))}. \item Instances of these classes can be decomposed through \emph{pattern matching} as we see below. \end{itemize} These features enable us to use Scala case classes much like we would use algebraic data types in a functional language. To explore the use of case classes, consider a function to evaluate an expression in some \emph{environment}. The purpose of an environment is to associate values with variables. For example, the expression \texttt{x+1} evaluated in an environment that associates the value \texttt{5} with the variable \texttt{x}, written \texttt{\{\ x\ -\textgreater{}\ 5\ \}}, gives \texttt{6} as result. An environment is just a function that associates a value with a (variable) name. The environment \texttt{\{\ x\ -\textgreater{}\ 5\ \}} given above can be written as a Scala function as follows: \begin{Shaded} \begin{Highlighting}[] \NormalTok{ \{ }\KeywordTok{case} \StringTok{"x"}\NormalTok{ => }\DecValTok{5}\NormalTok{ \}} \end{Highlighting} \end{Shaded} This notation defines a function that, when given the string \texttt{"x"} as argument, returns the integer \texttt{5}, and, otherwise, fails and throws an exception. An environment is a function of type \texttt{String\ =\textgreater{}\ Int}. To simplify our evaluation program, we define the name \texttt{Environment} to be an alias for this type using the following declaration: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{type}\NormalTok{ Environment = String => Int } \end{Highlighting} \end{Shaded} We can now define the evaluation function in Scala as follows: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{def} \FunctionTok{eval}\NormalTok{(t: Tree, env: Environment): Int = t }\KeywordTok{match}\NormalTok{ \{} \KeywordTok{case} \FunctionTok{Sum}\NormalTok{(l, r) => }\FunctionTok{eval}\NormalTok{(l, env) + }\FunctionTok{eval}\NormalTok{(r, env) } \KeywordTok{case} \FunctionTok{Var}\NormalTok{(n) => }\FunctionTok{env}\NormalTok{(n) } \KeywordTok{case} \FunctionTok{Const}\NormalTok{(v) => v } \NormalTok{ \}} \end{Highlighting} \end{Shaded} This evaluation function performs \emph{pattern matching} on the tree \texttt{t} (i.e., using the \texttt{match} operator). \begin{enumerate} \def\labelenumi{\arabic{enumi}.} \item The pattern match first checks whether the tree \texttt{t} is a \texttt{Sum} object (i.e., an instance of the \texttt{Sum} case class). If the tree is a \texttt{Sum}, the pattern match binds the left subtree to a new variable \texttt{l} and the right subtree to a new variable \texttt{r}, then evaluates the expression following the arrow \texttt{=\textgreater{}}. This expression uses the values of the variables bound by the pattern match, i.e., \texttt{l} and \texttt{r} in the \texttt{Sum} case. \item If the first check does not succeed, that is, the tree is not a \texttt{Sum} object, the pattern match then checks whether \texttt{t} is a \texttt{Var} object. If the tree is a \texttt{Var}, the pattern match binds the name contained in the \texttt{Var} node to a new variable \texttt{n} and then evaluates the right-hand-side expression. \item If the second check also fails, that is, if \texttt{t} is neither a \texttt{Sum} nor a \texttt{Var}, the pattern match then checks whether the expression is a \texttt{Const} object. If the tree is a \texttt{Const}, the pattern match binds the value contained in the \texttt{Const} node to a new variable \texttt{v} and then evaluates the right-hand-side expression. \item Finally, if all checks fail, the pattern match expression raises an exception to signal failure. In this version of \texttt{eval}, failure occurs only when there are additional subclasses of \texttt{Tree} that we have declared but not yet defined in the \texttt{eval} pattern match. \end{enumerate} Pattern matching attempts to match a value to a series of patterns. As soon as a pattern matches (moving top to bottom, left to right in the source code), the program extracts and names various parts of the value and then evaluates the associated expression using the values of these named parts. An object-oriented programmer might ask why we did not define \texttt{eval} as a \emph{method} of class \texttt{Tree} and its subclasses. We could do that, because Scala allows method definitions in case classes just as in normal classes. Deciding whether to use pattern matching or methods is partly a matter of taste. But the choice also affects the extensibility the program. \begin{itemize} \item Using methods, we can easily add a new kind of node by defining a new subclass of \texttt{Tree}. But adding a new operation to manipulate the tree is tedious because it requires us to modify every subclass of \texttt{Tree}. \item Using pattern matching, the situation is reversed: adding a new kind of node requires us to modify all functions that do pattern matching on the tree, to take the new node into account. But adding a new operation is easy, we just define it as an independent function. \end{itemize} To explore pattern matching further, consider another operation on arithmetic expressions: symbolic derivation. Looking back at our calculus class, we see the following rules for differentiation: \begin{enumerate} \def\labelenumi{\arabic{enumi}.} \item The derivative of a sum is the sum of the derivatives. \item The derivative of some variable \texttt{v} is 1 if \texttt{v} is the variable relative to which the derivation takes place, and is 0 otherwise. \item The derivative of a constant is 0. \end{enumerate} We can directly translate these rules into a Scala function that uses the above case classes and pattern matching, as follows: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{def} \FunctionTok{derive}\NormalTok{(t: Tree, v: String): Tree = t }\KeywordTok{match}\NormalTok{ \{} \KeywordTok{case} \FunctionTok{Sum}\NormalTok{(l, r) => }\FunctionTok{Sum}\NormalTok{(}\FunctionTok{derive}\NormalTok{(l, v), }\FunctionTok{derive}\NormalTok{(r, v))} \KeywordTok{case} \FunctionTok{Var}\NormalTok{(n) }\KeywordTok{if}\NormalTok{ (v == n) => }\FunctionTok{Const}\NormalTok{(}\DecValTok{1}\NormalTok{)} \KeywordTok{case}\NormalTok{ _ => }\FunctionTok{Const}\NormalTok{(}\DecValTok{0}\NormalTok{)} \NormalTok{ \}} \end{Highlighting} \end{Shaded} Function \texttt{derive} introduces two new concepts related to pattern matching. \begin{enumerate} \def\labelenumi{\arabic{enumi}.} \item The \texttt{case} expression for variables has a \emph{guard}, an expression following the \texttt{if} keyword. This guard prevents pattern matching from succeeding unless its expression is true. Here the guard ensures that the function returns the constant \texttt{1} only if the name of the variable being derived is the same as the derivation variable \texttt{v}. \item A pattern can include a \emph{wildcard}, written \texttt{\_}, that matches any value, without giving it a name. \end{enumerate} Consider an example with a simple \texttt{main} function that performs several operations on the expression \texttt{(x+x)+(7+y)}. It first computes its value in the environment \texttt{\{\ x\ -\textgreater{}\ 5,\ y\ -\textgreater{}\ 7\ \}} and then computes its derivative relative to \texttt{x} and then to \texttt{y}. \begin{Shaded} \begin{Highlighting}[] \KeywordTok{def} \FunctionTok{main}\NormalTok{(args: Array[String]) \{} \KeywordTok{val}\NormalTok{ exp: Tree = }\FunctionTok{Sum}\NormalTok{(}\FunctionTok{Sum}\NormalTok{(}\FunctionTok{Var}\NormalTok{(}\StringTok{"x"}\NormalTok{),}\FunctionTok{Var}\NormalTok{(}\StringTok{"x"}\NormalTok{)),}\FunctionTok{Sum}\NormalTok{(}\FunctionTok{Const}\NormalTok{(}\DecValTok{7}\NormalTok{),}\FunctionTok{Var}\NormalTok{(}\StringTok{"y"}\NormalTok{)))} \KeywordTok{val}\NormalTok{ env: Environment = \{ }\KeywordTok{case} \StringTok{"x"}\NormalTok{ => }\DecValTok{5} \KeywordTok{case} \StringTok{"y"}\NormalTok{ => }\DecValTok{7}\NormalTok{ \}} \FunctionTok{println}\NormalTok{(}\StringTok{"Expression: "}\NormalTok{ + exp)} \FunctionTok{println}\NormalTok{(}\StringTok{"Evaluation with x=5, y=7: "}\NormalTok{ + }\FunctionTok{eval}\NormalTok{(exp, env))} \FunctionTok{println}\NormalTok{(}\StringTok{"Derivative relative to x:}\CharTok{\textbackslash{}n}\StringTok{ "}\NormalTok{ + }\FunctionTok{derive}\NormalTok{(exp, }\StringTok{"x"}\NormalTok{))} \FunctionTok{println}\NormalTok{(}\StringTok{"Derivative relative to y:}\CharTok{\textbackslash{}n}\StringTok{ "}\NormalTok{ + }\FunctionTok{derive}\NormalTok{(exp, }\StringTok{"y"}\NormalTok{))} \NormalTok{ \}} \end{Highlighting} \end{Shaded} Executing this program, we get the expected output: \begin{verbatim} Expression: Sum(Sum(Var(x),Var(x)),Sum(Const(7),Var(y))) Evaluation with x=5, y=7: 24 Derivative relative to x: Sum(Sum(Const(1),Const(1)),Sum(Const(0),Const(0))) Derivative relative to y: Sum(Sum(Const(0),Const(0)),Sum(Const(0),Const(1))) \end{verbatim} The result of the derivative is complex. It should be simplified before printing. Defining a basic simplification function using pattern matching is an interesting (but surprisingly tricky) problem, left as an exercise for the reader. \hypertarget{expression-tree-code}{% \subsubsection{Expression Tree Code}\label{expression-tree-code}} The Scala source code for the case class version of the Expression Tree program is in file \href{ExprCase.scala}{\texttt{ExprCase.scala}}. A traditional object-oriented version of the Expression Tree program is in file \href{ExprObj.scala}{\texttt{ExprObj.scala}}. \hypertarget{traits}{% \subsection{Traits}\label{traits}} In addition to inheriting code from a superclass, a Scala class can also reuse code from one or several \emph{traits}. From a Java perpective, traits can be viewed as interfaces that can also contain code. In Scala, when a class inherits from a trait, it implements that trait's interface, and inherits all the code contained in the trait. Note: In older Java implementations, an interface consisted of just method signatures and constants. In Java 8, interfaces can now include default definitions of methods, somewhat similar to Scala traits. To see the usefulness of traits, consider a classic example: ordered objects. We want to compare objects of a given class among themselves. For example, we need to compare objects according to some total order to sort them. In Java, objects that can be compared implement the \texttt{Comparable} interface. In Scala, we can do better by defining the equivalent of \texttt{Comparable} as a trait, which we call \texttt{Ord}. When comparing objects, six different comparison operations are useful: smaller, smaller or equal, equal, not equal, greater or equal, and greater. But it is tedious to define all of these for every class whose instances we wish to compare. We observe that we can define four of the six in terms of the other two. For example, given the equal and smaller comparison operators, we can define the other four comparison operators. In Scala, we can capture this observation in the following trait declaration: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{trait}\NormalTok{ Ord \{} \KeywordTok{def}\NormalTok{ < (that: Any): Boolean} \KeywordTok{def}\NormalTok{ <=(that: Any): Boolean = (}\KeywordTok{this}\NormalTok{ < that) || (}\KeywordTok{this}\NormalTok{ == that)} \KeywordTok{def}\NormalTok{ > (that: Any): Boolean = !(}\KeywordTok{this}\NormalTok{ <= that)} \KeywordTok{def}\NormalTok{ >=(that: Any): Boolean = !(}\KeywordTok{this}\NormalTok{ < that)} \NormalTok{ \}} \end{Highlighting} \end{Shaded} This definition both creates a new type called \texttt{Ord}, which plays the same role as Java's \texttt{Comparable} interface, and generates default implementations of three comparison operators in terms of a fourth, abstract operator. All classes inherit the equality and inequality operators and, thus, those operators do not need to be defined in \texttt{Ord}. The type \texttt{Any} used above, is the supertype of all other types in Scala. It is essentially a more general version of Java's \texttt{Object} type that is also a supertype of basic types like \texttt{Int}, \texttt{Float}, etc. To make objects of a class comparable, it is sufficient to define equality and inferiority operators and then ``mix in'' the \texttt{Ord} trait. For example, consider a \texttt{Date} class representing dates in the Gregorian calendar. Such dates are composed of a day, a month, and a year, each of which we can represent with an integer. We can start the definition of the \texttt{Date} class as follows: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{class}\NormalTok{ Date(y: Int, m: Int, d: Int) }\KeywordTok{extends}\NormalTok{ Ord \{} \KeywordTok{def}\NormalTok{ year = y} \KeywordTok{def}\NormalTok{ month = m} \KeywordTok{def}\NormalTok{ day = d} \KeywordTok{override} \KeywordTok{def} \FunctionTok{toString}\NormalTok{(): String = year + }\StringTok{"-"}\NormalTok{ + month + }\StringTok{"-"}\NormalTok{ + day} \CommentTok{// ... two methods defined below} \NormalTok{ \}} \end{Highlighting} \end{Shaded} The \texttt{extends\ Ord} declaration specifies that the \texttt{Date} class inherits from the \texttt{Ord} trait (in addition to the default superclass). To make the comparisons work correctly, we redefine the \texttt{equals} method, inherited from Java's \texttt{Object} class, to compare dates by comparing their individual fields. The default implementation of \texttt{equals} does not work because it compares objects physically. In Scala, the comparison becomes: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{override} \KeywordTok{def} \FunctionTok{equals}\NormalTok{(that: Any): Boolean =} \NormalTok{ that.}\FunctionTok{isInstanceOf}\NormalTok{[Date] && \{} \KeywordTok{val}\NormalTok{ o = that.}\FunctionTok{asInstanceOf}\NormalTok{[Date]} \NormalTok{ o.}\FunctionTok{day}\NormalTok{ == day && o.}\FunctionTok{month}\NormalTok{ == month && o.}\FunctionTok{year}\NormalTok{ == year} \NormalTok{ \}} \end{Highlighting} \end{Shaded} This method uses the predefined methods \texttt{isInstanceOf} and \texttt{asInstanceOf}. \begin{itemize} \item \texttt{isInstanceOf} corresponds to the Java \texttt{instanceof} operator. It returns true if and only if the object to which it is applied is an instance of the given type. \item \texttt{asInstanceOf} corresponds to the Java cast operator: if the object is an instance of the given type, it is viewed as such, otherwise it throws a \texttt{ClassCastException}. \end{itemize} To complete our implementation, we need to define the \texttt{\textless{}} operator. It uses the predefined Scala method \texttt{error} to throw an exception with the given error message. \begin{Shaded} \begin{Highlighting}[] \KeywordTok{def}\NormalTok{ <(that: Any): Boolean = \{} \KeywordTok{if}\NormalTok{ (!that.}\FunctionTok{isInstanceOf}\NormalTok{[Date])} \FunctionTok{error}\NormalTok{(}\StringTok{"cannot compare "}\NormalTok{ + that + }\StringTok{" and a Date"}\NormalTok{)} \KeywordTok{val}\NormalTok{ o = that.}\FunctionTok{asInstanceOf}\NormalTok{[Date]} \NormalTok{ (year < o.}\FunctionTok{year}\NormalTok{) ||} \NormalTok{ (year == o.}\FunctionTok{year}\NormalTok{ && (month < o.}\FunctionTok{month}\NormalTok{ ||} \NormalTok{ (month == o.}\FunctionTok{month}\NormalTok{ && day < o.}\FunctionTok{day}\NormalTok{)))} \NormalTok{ \}} \end{Highlighting} \end{Shaded} This completes the definition of the \texttt{Date} class. Instances of the \texttt{Date} class can be seen either as dates or as comparable objects. They define all six comparison operators: \begin{itemize} \item \texttt{equals} and \texttt{\textless{}} because they appear directly in the definition of the \texttt{Date} class \item the other four because they are inherited from the \texttt{Ord} trait \end{itemize} Traits are useful in many other situations, as you will see as you program more in Scala. \hypertarget{genericity}{% \subsection{Genericity}\label{genericity}} Scala supports generics. Java did not support generics until Java 5. \emph{Genericity} is the ability to write code parameterized by types. For example, suppose we are writing a library for linked lists. What type do we give the elements of the list? \begin{itemize} \item If we choose a specific concrete type such as \texttt{Int}, we severely limit the usefulness of the library. It is impractical to include different implementation for every possible concrete type. \item If follow choose a default supertype like \texttt{Any}, then users of our library would need to their lace code with many type checks and type casts. (In Java, there was also the problem that basic types do not inherit from \texttt{Object}.) \end{itemize} Scala supports generic classes and methods to solve this problem. As an example, consider the simplest possible container class: a reference, which can either be empty or point to an object of some type. \begin{Shaded} \begin{Highlighting}[] \KeywordTok{class}\NormalTok{ Reference[T] \{} \KeywordTok{private} \KeywordTok{var}\NormalTok{ contents: T = _} \KeywordTok{def} \FunctionTok{set}\NormalTok{(value: T) \{ contents = value \}} \KeywordTok{def}\NormalTok{ get: T = contents} \NormalTok{ \}} \end{Highlighting} \end{Shaded} This example parameterizes the class \texttt{Reference} with a type \texttt{T}, which is the type of its element. The body of the class declares the \texttt{contents} variable, the argument of the \texttt{set} method, and the return type of the \texttt{get} method to have type `T'. This is our first example to use mutable variables declare with \texttt{var}. In this example, we initialize the \texttt{contents} variable to have the value \texttt{\_}, which represents the default value for the type. This default value is 0 for numeric types, \texttt{false} for the \texttt{Boolean} type, \texttt{()} for the \texttt{Unit} type, and \texttt{null} for all object types. To use this \texttt{Reference} class, we need to specify what type to use for the type parameter \texttt{T}, that is, the type of the element contained in the cell. For example, to create and use a cell holding an integer, one could write the following: \begin{Shaded} \begin{Highlighting}[] \KeywordTok{object}\NormalTok{ IntegerReference \{} \KeywordTok{def} \FunctionTok{main}\NormalTok{(args: Array[String]) \{} \KeywordTok{val}\NormalTok{ cell = }\KeywordTok{new}\NormalTok{ Reference[Int]} \NormalTok{ cell.}\FunctionTok{set}\NormalTok{(}\DecValTok{13}\NormalTok{)} \FunctionTok{println}\NormalTok{(}\StringTok{"Reference contains the half of "}\NormalTok{ + (cell.}\FunctionTok{get}\NormalTok{ * }\DecValTok{2}\NormalTok{))} \NormalTok{ \}} \NormalTok{ \}} \end{Highlighting} \end{Shaded} This example does not need to cast the value returned by the \texttt{get} method before using it as an integer. It is also not possible to store anything but an integer in that particular cell, because it was declared as holding an integer. \hypertarget{conclusion}{% \subsection{Conclusion}\label{conclusion}} The goal of this document is to provide a quick overview of Scala for programmers already familiar with Java. To learn more, use a textbook, reference book, other tutorials, and manuals. \hypertarget{exercises}{% \subsection{Exercises}\label{exercises}} TODO: More exercises \begin{enumerate} \def\labelenumi{\arabic{enumi}.} \item Extend the \href{ExprCase.scala}{case class version of the expression-tree calculator program} as described below. \begin{enumerate} \def\labelenumii{\alph{enumii}.} \item Modify the \texttt{Tree} algebraic data type as follows: \begin{itemize} \item Change node type \texttt{Const} to represent a floating point number instead of an integer. \item Add node types \texttt{Sub}, \texttt{Prod}, and \texttt{Div} to represent subtraction, multiplication, and division of the two values, respectively. (Assume \texttt{Sub} means that the right operand is subtracted from the left operand.) \item add node type \texttt{Neg} to represent negating a value. \item add node types \texttt{Sin} and \texttt{Cos} to represent the sine and cosine trigonometric functions, respectively. \end{itemize} \item Extend functions \texttt{eval} and \texttt{derive} to support the above modifications of \texttt{Tree}. Test the extended functions thoroughly. (What should be done for division by zero?) \end{enumerate} \item Further extend the expression \texttt{Tree} program developed in the previous exercise as described below. \begin{enumerate} \def\labelenumii{\alph{enumii}.} \item Develop a new function \texttt{simplify} that takes an extended expression \texttt{Tree}, simplifies it by evaluating \emph{all} subexpressions involving only constants (i.e., not evaluating variables), and returns the new expression \texttt{Tree}. For example: \begin{itemize} \tightlist \item \texttt{simplify(Add(Const(2),Const(3)))} yields \texttt{Const(5)} \item \texttt{simplify(Add(Add(Const(1),Const(1)),Mul(Const(1),Const(3))))} yields \texttt{Const(5)}. \item \texttt{simplify(Neg(Const(5)))} yields \texttt{Const(-5)} \end{itemize} \item Extend the \texttt{simplify} function to exploit mathematical properties such as identity elements (\texttt{x\ +\ 0\ =\ x\ =\ 0\ +\ x}), and zeros \item Further extend the \texttt{simplify} function to exploit associativity (\texttt{(x\ +\ y)\ +\ z\ =\ x\ +\ (y\ +\ z)}, commutativity (\texttt{x\ +\ 1\ =\ 1\ +\ x}), etc. \end{enumerate} \item Modify and extend the \href{ExprObj.scala}{traditional object-oriented version of the expression-tree calculator program} in the same manner as described in the previous two exercises. \end{enumerate} \hypertarget{acknowledgements}{% \subsection{Acknowledgements}\label{acknowledgements}} In February 2016, I created these notes by adapting and expanding the web document ``A Scala Tutorial for Java Programmers'' by Michel Schinz and Phillipp Haller to better meet the needs of my Scala-based course (to explain some items more thoroughly, improve the narrative, and make the document accessible). The original document can be found on the Scala Language website at \url{http://docs.scala-lang.org/tutorials/scala-for-java-programmers.html}. In Spring 2017, I updated the format to be better compatible with Pandoc. I revised the document further in Spring 2018 (e.g., adding the first exercises). I maintain these notes as text in Pandoc's dialect of Markdown using embedded LaTeX markup for the mathematical formulas and then translate the notes to HTML, PDF, and other forms as needed. \hypertarget{references}{% \subsection{References}\label{references}} TODO: Add \hypertarget{concepts}{% \subsection{Concepts}\label{concepts}} TODO: Add \end{document}