Today. cis15-spring2010-parsons-lectii.1 2

Transcription

1 CLASSES AND OBJECTS

2 Today We will start to talk about object-oriented programming In particular we will talk about struct and class. We willshow how touse these features of C++to define aggregate data types. We willshow how todefine methods thatoperate onthese data types. Thiswork isbased onpohl, Chapter 4. Muchoftheworkwewilldoforthenextcouple ofweekswillbe concerned notonly with whatwecan doinc++, butalso the style inwhich wedoit. cis15-spring2010-parsons-lectii.1 2

3 Aggregate data types New today: class and struct struct comes from C class isnew inc++ andyou shouldhave learnt about itincis 1.5. Both are aggregate types, meaning that they group together multiple fields of data. Forexample wehave: class point { public: double x, y; }; cis15-spring2010-parsons-lectii.1 3

4 We can alsowrite: struct s_point { double x, y; }; Don t forget toputasemi-colon at theend of the structure definition! cis15-spring2010-parsons-lectii.1 4

5 Aside: why is point useful? The idea behind point is that it represents information about the location of something. Think of itasapair of(cartesian) coordinates. We group the coordinates together because they make no sense separately if wehave the x coordinate of athing,then ithasa y coordinate also. We willuse point when wewriteasimulation of small eco-system and ofarobot operating inasimulated world. We willdo thisin someofthe homeworks. cis15-spring2010-parsons-lectii.1 5

6 Back to aggregate data types In C,the tag (point) isoptional and does notconstitute adata type (you need to use typedef as well). In C++, thetag is considered adatatype, hence the above example is a data type definition. Thismeans that youcan use point asadatatype, e.g.: point p; In other words,you can declare avariable p whichisof type point. p iscalled an objectoran instanceofclass point. cis15-spring2010-parsons-lectii.1 6

7 The fields orelements ofan aggregatedata typeare called members. Members are referred to using dot notation, e.g.: p.x = 7.0; p.y = 10.3; You can alsouse apointerto access members of anaggregate data type, e.g.: p->x = 12.3; but wewilldiscusspointers inthe next unit,sodon t worry about thisnow... cis15-spring2010-parsons-lectii.1 7

8 Just asyou candefine anobject oftype point: point p; you can define an array ofthese objects point mypoints[3]; and even point mypoints[3] = {{1, 2}, {3, 4}, {5, 6}}; which defines the array mypoints to hold three elements each of which is a class of type point which holds two doubles, and sets the valuesof these. We can then access the individual members asbefore: cout << mypoints[1].x; will, for example, print out 3. cis15-spring2010-parsons-lectii.1 8

9 Member functions In C++, members ofaggregate datatypes can be functions (C only allowsdata members) In object-oriented programming (OOP) lingo, the word method is often used instead of function The reasonto define functions inside anaggregate datatype isto follow the OOP principle of encapsulation operations should be packaged with data Thisisastylething. For example: cis15-spring2010-parsons-lectii.1 9

10 #include <iostream> using namespace std; class point { public: double x, y; void print() { cout << "(" << x << "," << y << ")\n"; } void set( double u, double v ) { x = u; y = v; } }; // end of class--don t forget semi-colon! int main() { point w; w.set( 1.2, 3.4 ); cout << "point = "; w.print(); } cis15-spring2010-parsons-lectii.1 10

11 Notes: Notice that the set method changes the values ofthe data members this is considered good OOP practise Defining the methods inside the class definition is called in-line declaration ; this is generally only okay for short, concise methods The class scope operator can be used when in-line declarations are inappropriate. For example: cis15-spring2010-parsons-lectii.1 11

12 #include <iostream> using namespace std; class point { public: double x, y; void print(); void set( double u, double v ); }; void point::print() { cout << "(" << x << "," << y << ")\n"; } // end of print() void point::set( double u, double v ) { x = u; y = v; } // end of set() cis15-spring2010-parsons-lectii.1 12

13 The methods can then beinvoked from main. Asfordatamembers, weinvokefunction membersusingthe dot notation. int main() { point w; w.set( 1.2, 3.4 ); cout << "point = "; w.print(); } // end of main() cis15-spring2010-parsons-lectii.1 13

14 Class scope The classscope operator istwocolons (::), as inourexample: void point::print() const { cout << "(" << x << "," << y << ")\n"; } The ::operator hasthe highest precedence in thelanguage, soit always gets evaluated first There are twoversions ofthe operator: binary andunary The binary version isthe one weused before: point::print(), which is used to refer to a variable s class scope (also called local scope ). The unary version islikethis: ::count andisused torefer toa variable s external scope (e.g., for a global variable). cis15-spring2010-parsons-lectii.1 14

15 Here isa(maybe confusing) example fromthe book: int count = 0; // declare global variable void how_many( double w[], double x, int& count ) { for ( int i=0; i<n; ++i ) { count += ( w[i] == x ); // local count } ++::count; // global count } // end of how_many() We need tousethe unary scope operator her since count is declared twice If youdidn t have the ::count, then thesecond time, the useof count would also refer to the local variable It isbetter practise notto useglobalvariables; oratleast if you do, give them unique names to avoid confusion :-) cis15-spring2010-parsons-lectii.1 15

16 this pointer The keyword this isusedto refer toaninstance of aclassfrom within itself. It isapointer something wewilldiscussat length inthe next unit Here isapossible use togive youthe idea. The data members are available anywhere inside any function members: point::foo(double a) { if(x == a){ cout << y; } cis15-spring2010-parsons-lectii.1 16

17 But whatdoes x refer toin: point::bar(double x) { if(x == x){ cout << y; } Turnsoutisisthe xthat isthe argumenttothe function. Torefer to the xthatis the datamember use this: point::bar(double x) { if(this->x == x){ cout << y; } Thislast version of bar isthe same as foo. cis15-spring2010-parsons-lectii.1 17

18 Summary This lecture introduced the basics of object-oriented programming. It showedhow struct and class can be usedtocreate aggregate datatypes. We looked at creating objects, instances of classes. And welooked atattaching methods toclasses. These are the main concepts you need for object-oriented programming. Everything else is about making programs better (not more functional). cis15-spring2010-parsons-lectii.1 18