1.

SARASWATI COMPUTER CENTRE

Welcome to Saraswati Education Centre
Saraswati Education Centre was established on 31st of July 2003 vide Jammu & Kashmir
ovt! "e#istration $o! %C&11&12'(' )ated 31&0*&2003! +his ,nstitute is an erstwhile
,nstitute of Com-uter and .cademic Education and en/oys -rivile#e of rich e0-erience in
the field of Education! 1ur ob/ectives are 2
+o 3rovide access to 4uality education to all those who see5 it6
irres-ective of a#e6 re#ion or formal 4ualifications!
"ela0ed entry re#ulations!
7le0ibility in choosin# the courses from wide ran#e of disci-lines!
)emocrati8in# education by ma5in# to the masses!
1
Accreditation
S.".SW.+, E)9C.+,1$ CE$+"E has #ot accredited with KS19 for runnin#
9niversity courses of :ana#ement6 ,nformation +echnolo#y 3ro#rams and Com-uter
.--lication 3ro#rams since )ec 200; for the benefit of all as-irants6 lobal 1-en
9niversity6 $a#aland from June 200(6 <hartiya Shi5sha 3arishad and Sin#hania
9niversity as 3lacement and Counsellin# centre for various courses! S.".SW.+,
E)9C.+,1$ CE$+"E is also certified by ,S1 =00122000 Certification as 4uality and
standards followed by the or#ani8ation!
2
Courses of Saraswati Computer Centre :
COURSES DURATION FEES
1ffice 1-erations 3 months 1;00&>
C&C?? 2 month 2;00&>
%isual <asic 2 months 2;00&>
[email protected] and .--licationsA 2 months 3000&>
7o0 3ro 2 months 1;00&>
%isual C?? 2 months 2;00&>
!$et 3ro#rammin# ' months ;000&>
@%isual ! net 6 CB6 .S3 !$etA
.ccountin# 3ac5a#e 3 months 3C00&>
@:anual .&c6 +ally6 <usyA
+ally 1 month 2000&>
<usy 1 month 2000&>
E0>$e0t eneration 1 month 2000&>
)+3 2 months 1(00&>
@3a#e:a5er6 Corel )rawA
3hotosho- 1 month (00&>
1racle 1 1&2 month 1;00&>
D+:E@Web )esi#nin#A 2 months 2;00&>
@D+:E6 F:E6 7ront3a#eA
%<Scri-t&JavaScri-t 1 month 1000&>
.S3 1 1&2 month 1;00&>
3
'
2. C++
C?? @read as GC 3lus 3lusGA is a statically ty-ed6 free>form6 multi>-aradi#m6 com-iled6
#eneral>-ur-ose -ro#rammin# lan#ua#e! ,t is re#arded as a middle>level lan#ua#e6 as it
com-rises a combination of both hi#h>level and low>level lan#ua#e features! ,t was
develo-ed by </arne Stroustru- startin# in 1=*= at <ell Eabs as an enhancement to the C
-ro#rammin# lan#ua#e and ori#inally named GC with ClassesG! ,t was renamed to C?? in
1=(3!
C?? is widely used in the software industry6 and remains one of the most -o-ular
lan#ua#es ever created! Some of its a--lication domains include systems software6
a--lication software6 device drivers6 embedded software6 hi#h>-erformance server and
client a--lications6 and entertainment software such as video #ames! Several #rou-s
-rovide both free and -ro-rietary C?? com-iler software6 includin# the $9 3ro/ect6
:icrosoft6 ,ntel6 <orland and others!
+he lan#ua#e be#an as enhancements to C6 first addin# classes6 then virtual functions6
o-erator overloadin#6 multi-le inheritance6 tem-lates6 and e0ce-tion handlin# amon#
other features! .fter years of develo-ment6 the C?? -ro#rammin# lan#ua#e standard was
ratified in 1==( as ,S1&,EC 1'((221==(! +hat standard is still current6 but is amended by
the 2003 technical corri#endum6 ,S1&,EC 1'((222003! +he ne0t standard version @5nown
informally as C??00A is in develo-ment!
+odays com-uters are incredibly fast6 and #ettin# faster all the time! Het with this s-eed
comes some si#nificant constraints! Com-uters only natively understand a very limited
set of instructions6 and must be told e0actly what to do! +he set of instructions that tells a
com-uter what to do is 5nown as software! +he com-uter machinery that e0ecutes the
instructions is the hardware!
. com-uterIs C39 is inca-able of s-ea5in# C??! +he very limited set of instructions that
a C39 natively understands is called machine code6 or machine lan#ua#e6 or an
instruction set! Dow these instructions are or#ani8ed is beyond the sco-e of this
;
introduction6 but it is interestin# to note two thin#s! 7irst6 each instruction is com-osed of
a number of binary di#its6 each of which can only be a 0 or a 1! +hese binary numbers are
often called bits @short for binary di#itA! 7or e0am-le6 the :,3S architecture instruction
set always has instructions that are 32 bits lon#! 1ther architectures @such as the 0(C6
which you are li5ely usin#A have instructions that can be a variable len#th!
7or e0am-le6 here is a 0(C machine lan#ua#e instruction2 10110000 01100001
Second6 each set of binary di#its is translated by the C39 into an instruction that tells it to
do a very s-ecific /ob6 such as com-are these two numbers6 or -ut this number in that
memory location! )ifferent ty-es of C39s will ty-ically have different instruction sets6
so instructions that would run on a 3entium ' would not run on a :acintosh 3ower3C
based com-uter! <ac5 when com-uters were first invented6 -ro#rammers had to write
-ro#rams directly in machine lan#ua#e6 which was a very difficult and time consumin#
thin# to do!
<ecause machine lan#ua#e is so hard to -ro#ram with6 assembly lan#ua#e was invented!
,n an assembly lan#ua#e6 each instruction is identified by a short name @rather than a set
of bitsA6 and variables can be identified by names rather than numbers! +his ma5es them
much easier to read and write! Dowever6 the C39 can not understand assembly lan#ua#e
directly! ,nstead6 it must be translated into machine lan#ua#e by usin# an assembler!
.ssembly lan#ua#es tend to be very fast6 and assembly is still used today when s-eed is
critical! Dowever6 the reason assembly lan#ua#e is so fast is because assembly lan#ua#e
is tailored to a -articular C39! .ssembly -ro#rams written for one C39 will not run on
another C39! 7urthermore6 assembly lan#ua#es still re4uire a lot of instructions to do
even sim-le tas5s6 and are not very human readable!
Dere is the same instruction as above in assembly lan#ua#e2 mov al6 0C1h
+o address these concerns6 hi#h>level -ro#rammin# lan#ua#es were develo-ed! C6 C??6
3ascal6 .da6 Java6 Javascri-t6 and 3erl6 are all hi#h level lan#ua#es! 3ro#rams written in
C
hi#h level lan#ua#es must be translated into a form that the C39 can understand before
they can be e0ecuted! +here are two -rimary ways this is done2 com-ilin# and
inter-retin#!
. com-iler is a -ro#ram that reads code and -roduces a stand>alone e0ecutable that the
C39 can understand directly! 1nce your code has been turned into an e0ecutable6 you do
not need the com-iler to run the -ro#ram! .lthou#h it may intuitively seem li5e hi#h>
level lan#ua#es would be si#nificantly less efficient than assembly lan#ua#es6 modern
com-ilers do an e0cellent /ob of convertin# hi#h>level lan#ua#es into fast e0ecutables!
Sometimes6 they even do a better /ob than human coders can do in assembly lan#ua#eJ
Dere is a sim-lified re-resentation of the com-ilin# -rocess2
Fig. 2.1 : Co!i"ing Proce##
.n inter-reter is a -ro#ram that reads code and essentially com-iles and e0ecutes
@inter-retsA your -ro#ram as it is run! 1ne advanta#e of inter-reters is that they are much
easier to write than com-ilers6 because they can be written in a hi#h>level lan#ua#e
themselves! Dowever6 they tend to be less efficient when runnin# -ro#rams because the
com-ilin# needs to be done every time the -ro#ram is run! 7urthermore6 the inter-reter is
needed every time the -ro#ram is run!
Dere is a sim-lified re-resentation of the inter-retation -rocess2
Fig. 2.2 Inter!retation Proce##
*
.ny lan#ua#e can be com-iled or inter-reted6 however6 traditionally lan#ua#es li5e C6 C?
?6 and 3ascal are com-iled6 whereas Kscri-tin#L lan#ua#es li5e 3erl and Javascri-t are
inter-reted! Some lan#ua#es6 li5e Java6 use a mi0 of the two!
Di#h level lan#ua#es have several desirable -ro-erties! 7irst6 hi#h level lan#ua#es are
much easier to read and write!
Dere is the same instruction as above in C&C??2 a M =*N
Second6 they re4uire less instructions to -erform the same tas5 as lower level lan#ua#es!
,n C?? you can do somethin# li5e a M b O 2 ? ;N in one line! ,n assembly lan#ua#e6 this
would ta5e ; or C different instructions!
+hird6 you donIt have to concern yourself with details such as loadin# variables into C39
re#isters! +he com-iler or inter-reter ta5es care of all those details for you! .nd fourth6
they are -ortable to different architectures6 with one ma/or e0ce-tion6 which we will
discuss in a moment!
Fig. 2.$ : Arc%itect&re
+he e0ce-tion to -ortability is that many -latforms6 such as :icrosoft Windows6 contain
-latform>s-ecific functions that you can use in your code! +hese can ma5e it much easier
to write a -ro#ram for a s-ecific -latform6 but at the e0-ense of -ortability! ,n these
tutorials6 we will e0-licitly -oint out whenever we show you anythin# that is -latform
s-ecific!
(
$. A FEW COMMON C++ PRO'(EMS
,n this section6 weIll address some of the common issues that new -ro#rammers seem to
run across with fairly hi#h -robability! +his is not meant to be a com-rehensive list of
com-ilation or e0ecution -roblems6 but rather a -ra#matic list of solutions to very basic
issues! ,f you have any su##estions for other issues that mi#ht be added to this list6 -ost
them in the comments section below!
3roblem 12 When e0ecutin# a -ro#ram from the ,)E6 the console window blin5s and then
closes immediately!
.nswer 12 Some com-ilers @e#! <loodshedIs )ev C??A donIt automatically -ause the
console screen after the -ro#ram has finished e0ecutin#! ,f this is the case with your
com-iler6 the followin# two ste-s will fi0 your -roblem2
7irst6 add the followin# line near the to- of your -ro#ram2
Binclude PiostreamQ
Second6 add the followin# code at the end of the main@A function @ri#ht before the return
statementA2
cin!clear@AN
cin!i#nore@2;;6 RSnRAN
cin!#et@AN
+his will cause your -ro#ram to wait for you to -ress a 5ey before continuin#6 which will
#ive you time to e0amine your -ro#ramIs out-ut before your com-iler closes the console
window!
1ther solutions6 such as the commonly su##ested system@L-auseLA solution may only
wor5 on certain o-eratin# systems!
=
3roblem 22 When com-ilin# with :icrosoft %isual C??6 you #et the followin# error2
Kc2Svc-ro/ectsStest!c--@2C3A 2fatal error C10102 une0-ected end of file while loo5in# for
-recom-iled header directiveL
.nswer 22 +his error occurs when the :icrosoft %isual C?? com-iler is set to use
-recom-iled headers but one of your C?? code files does not include the stdaf0 header as
the first line! +o fi0 this -roblem6 sim-ly locate the file -roducin# the error @in the above
error6 test!c-- is the cul-ritA6 and add the followin# line at the very to- of the file2
Binclude Gstdaf0!hG
.lternatively6 you can turn off -recom-iled headers!
3roblem 32 When tryin# to use cin or cout6 the com-iler says cin or cout is an
Kundeclared identifierL
.nswer 32 7irst6 ma5e sure you have included the followin# line near the to- of your file2
Binclude PiostreamQ
,n any function where you use use cin or cout6 include the followin# line2
usin# names-ace stdN
3roblem '2 When tryin# to use endl to end a -rinted line6 the com-iler says end1 is an
Kundeclared identifierL
:a5e sure you do not mista5e the letter l @lower case EA in endl for the number 1! endl is
all letters! , recommend usin# a font that ma5es it clear the differences between the letter
lower case E6 u--er case i6 and the number 1! .lso the letter ca-ital o and the number
8ero can easily be confused in many fonts! +here is a hu#e list of -ro#rammin# fonts
here!
3roblem '2 :y -ro#ram com-iles but it isnIt wor5in# correctly! What do , doT
10
.nswer '2 )ebu# itJ Hou can find information on how to debu# -ro#rams in a--endi0 .6
s-ecifically sections .!' and .!;! Hou will -robably find the debu##in# sections more
com-rehensible after readin# a few more cha-ters6 but , wanted to ma5e you aware of
their e0istence now for future reference!
1!1 U Structure of a -ro#ram
,nde0
0!C U Com-ilin# your first -ro#ram
3rint +his 3ost
+his entry is filed under C?? +utorial! Hou can follow any res-onses to this entry throu#h
the "SS 2!0 feed! Hou can leave a res-onse6 or trac5bac5 from your own site!
;1 "es-onses to K0!* U . few common C?? -roblemsL
Comment by Jason
200*>12>1* 02222222
3roblem 12 When e0ecutin# a -ro#ram from the ,)E6 the console window blin5s and then
closes immediately!
+he ide ,Im usin# is %isual C !$E+ 2002! +his ide has -roblem 1!
+o fi0 the -roblem , include the followin# command at the end of the main@A function
before the return statement!
system@L-auseLAN
V 9nfortunately6 this only wor5s on the Windows o-eratin# system! +he system@A function
tells the o-eratin# system to e0ecute whatever the -arameter is >> in this case6 a command
named G-auseG! 1n Windows6 this waits for user in-ut! 1n other o-eratin# systems @such
as linu0A6 the -ause command does somethin# else entirely! Dowever6 if you are only
11
-ro#rammin# on Windows and not distributin# your source code6 you may find this an
easier solution while learnin#! >.le0 W
"e-ly to this comment

Comment by fo4
200*>12>1* 0*22;2'=
.nother #reat way to #et a -ro#ram to -ause is to re4uire a user in-ut @cinA! 9sin# a
menu>ty-e interface that re4uires the user to select an o-tion for Ke0itin#L wor5s well!
,Im only a be#inner6 but for sim-le calculators that ,Ive created6 this wor5s well!
"e-ly to this comment

Comment by Jason
200*>12>1( 1'2102''
,n order to #et the screen to -ause before e0itin# the -ro#ram6 does the code2
cin!clear@AN
cin!i#nore@2;;6 RSnRAN
cin!#et@AN
have any advanta#es over /ust usin#
cin QQ my%ariableN
, have so much to learn! 2A
"e-ly to this comment

Comment by .le0 @adminA
200*>12>1( 1'22(2'1
Jason6 #ood 4uestion! DereIs a sim-le -ro#ram that illustrates why cin QQ my%ariable is
not the best way to -ause2
12
Binclude PiostreamQ
int main@A
X
usin# names-ace stdN
&& .llocate a buffer to hold user in-ut
char bufV2;;W M X 0 YN
&& We only want one letter
cout PP GEnter a letter2 GN
cin QQ bufV0WN
&& 3ause and wait for user in-ut
char ch,#noreN
cout PP G3ausin#G PP endlN
cin QQ ch,#noreN
return 0N
Y
7irst6 cin QQ ch,#nore i#nores the enter 5ey if the user does not -rovide any other in-ut!
Hou can verify this by hittin# enter when the cin QQ ch,#nore line is bein# e0ecuted! 1n
the other hand6 cin!#et@A acce-ts the enter 5ey as valid in-ut!
<ut -erha-s more im-ortantly is the second -henomena that occurs when more characters
are -laced into the in-ut stream @cinA than are read out! When as5ed for a letter in the
above -ro#ram6 enter more than one letter and watch what ha--ens!
13
EetIs say you enter KabcdeL! cin QQ bufV0W reads the ZaI character6 but KbcdeL are still left
in the in-ut stream! Conse4uently6 when the code #ets to cin QQ ch,#nore6 it reads the
waitin# ZbI character out of the in-ut stream and doesnIt -ause at allJ
+he 3 ste- -rocess is a little more fool-roof! 7irst6 cin!clear@A clears any errors the user
may have caused by -rovidin# invalid in-ut in the -ast! Second6 cin!i#nore@2;;6 RnRA #ets
rid of u- to 2;; waitin# characters in the in-ut stream @or until a n is encounteredA6 which
is almost always #oin# to be enou#h to clear out the in-ut stream! 7inally6 we use
cin!#et@A to wait for a character! Since we /ust cleared the in-ut stream6 there are no
characters waitin#! Conse4uently6 the -ro#ram will wait for the user to enter somethin#!
,n ==[ of the cases you encounter6 cin QQ my%ariable will -robably wor5 e0actly as you
e0-ect! <ut since the 3 line method is #enerally more fool-roof and isnIt -articularly
burdensome to im-lement6 ,Id say itIs the su-erior solution!
1'
). STRUCTURE OF A PRO*RAMMIN*
. com-uter -ro#ram is a se4uence of instructions that tell the com-uter what to do!
Statements and expressions
+he most common ty-e of instruction in a -ro#ram is the statement! . statement in C??
is the smallest inde-endent unit in the lan#ua#e! ,n human lan#ua#e6 it is analo#ous to a
sentence! We write sentences in order to convey an idea! ,n C??6 we write statements in
order to convey to the com-iler that we want to -erform a tas5! Statements in C?? are
terminated by a semicolon!
+here are many different 5inds of statements in C??! +he followin# are some of the most
common ty-es of sim-le statements2
1.int x;
2.x = 5;
3.cout << x;
int 0 is a declaration statement! ,t tells the com-iler that 0 is a variable! .ll variables in a
-ro#ram must be declared before they are used! We will tal5 more about variables shortly!
0 M ; is an assi#nment statement! ,t assi#ns a value @;A to a variable @0A!
cout PP 0N is an out-ut statement! ,t out-uts the value of 0 @which we set to ; in the
-revious statementA to the screen!
+he com-iler is also ca-able of resolvin# e0-ressions! .n e0-ression is an mathematical
entity that evaluates to a value! 7or e0am-le6 in math6 the e0-ression 2?3 evaluates to the
value ;! E0-ressions can involve values @such as 2A6 variables @such as 0A6 o-erators @such
as ?A and functions @which return an out-ut value based on some in-ut valueA! +hey can
be sin#ular @such as 26 or 0A6 or com-ound @such as 2?36 2?06 0?y6 or @2?0AO@y>3AA!
7or e0am-le6 the statement 0 M 2 ? 3N is a valid assi#nment statement! +he e0-ression
2?3 evaluates to the value of ;! +his value of ; is then assi#ned to 0!
1;
Functions
,n C??6 statements are ty-ically #rou-ed into units called functions! . function is a
collection of statements that e0ecutes se4uentially! Every C?? -ro#ram must contain a
s-ecial function called main@A! When the C?? -ro#ram is run6 e0ecution starts with the
first statement inside of main@A! 7unctions are ty-ically written to do a very s-ecific /ob!
7or e0am-le6 a function named :a0@A mi#ht contain statements that fi#ures out which of
two numbers is lar#er! . function named Calculaterade@A mi#ht calculate a studentIs
#rade! We will tal5 more about functions later!
Libraries
Eibraries are #rou-s of functions that have been K-ac5a#ed u-L for reuse in many
different -ro#rams! +he core C?? lan#ua#e is actually very small and minimalistic U
however6 C?? comes with a bunch of libraries6 5nown as the C?? standard libraries6 that
-rovide -ro#rammers with lots of e0tra functionality! 7or e0am-le6 the iostream library
contains functions for doin# in-ut and out-ut! )urin# the lin5 sta#e of the com-ilation
-rocess6 the libraries from the C?? standard library are the runtime su--ort libraries that
are lin5ed into the -ro#ram @this will be discussed further in lesson 1!'A!
Taking a look at a sample program
$ow that you have a brief understandin# of what statements6 functions6 and libraries are6
letIs loo5 at a sim-le hello world -ro#ram!
Consider our hello world -ro#ram2
1.#include <iostream>
2.
3.int main()
4.{
5. using namespace std;
. cout << !"ello #orld$! << endl;
%. return &;
1C
'.(
Eine 1 is a s-ecial ty-e of statement called a -re-rocessor directive! 3re-rocessor
directives tell the com-iler to -erform a s-ecial tas5! ,n this case6 we are tellin# the
com-iler that we would li5e to use the iostream library! +he iostream library contains
code that tells the com-iler what cout and endl do! ,n other words6 we need to include the
iostream library in order to write to the screen!
Eine 3 declares the main@A function6 which as you learned above6 is mandatory! Every
-ro#ram must have a main@A function!
Eines ' and ( tell the com-iler which lines are -art of the main function! Everythin#
between the o-enin# curly brace on line ' and the closin# curly brace on line ( is
considered -art of the main@A function!
Eine ; is our first statement @you can tell itIs a statement because it ends with a
semicolonA! .s you learned in the e0-lanation for line 16 cout and endl live inside the
iostream library! Dowever6 within iostream6 they live inside a s-ecial com-artment named
std @short for standardA! +his usin# statement tells the com-iler to loo5 inside a
com-artment named std if it canIt find cout or endl defined anywhere else! ,n other
words6 this statement is also necessary so the com-iler can find cout and endl6 which we
use on line C!
Eine C is our out-ut statement! Cout is a s-ecial ob/ect that re-resents the console&screen!
+he PP symbol is an o-erator @much li5e ? is an o-eratorA called the out-ut o-erator!
Cout understands that anythin# sent to it via the PP o-erator should be -rinted on the
screen! Endl is a s-ecial symbol that moves the cursor to the ne0t line!
Eine * is a new ty-e of statement6 called a return statement! When an e0ecutable -ro#ram
finishes runnin#6 it sends a value to the o-eratin# system that indicates whether it was run
successfully or not! +he return value of main@A is used for this -ur-ose! +his -articular
return statement returns the value of 0 to the o-eratin# system6 which means Keverythin#
1*
went o5ayJL! $on>8ero numbers are ty-ically used to indicate that somethin# went
wron#6 and the -ro#ram had to abort! We will discuss return statements in more detail
when we discuss functions!
Conclusion
.ll of the -ro#rams we write will follow this tem-late6 or a variation on it! We will
discuss each of the lines above in more detail in the u-comin# sections!
Quiz
+he followin# 4ui8 is meant to reinforce your understandin# of the material -resented
above!
1A What is the difference between a statement and an e0-ressionT
2A What is the difference between a function and a libraryT
3A What symbol do statements in C?? end withT
1(
+. ,ISTOR-

</arne Stroustru-6 creator of C?? Stroustru- be#an wor5 on RC with ClassesR in 1=*=! +he
idea of creatin# a new lan#ua#e ori#inated from Stroustru-Rs e0-erience in -ro#rammin#
for his 3h!)! thesis! Stroustru- found that Simula had features that were very hel-ful for
lar#e software develo-ment6 but the lan#ua#e was too slow for -ractical use6 while <C3E
was fast but too low>level to be suitable for lar#e software develo-ment! When Stroustru-
started wor5in# in .+&+ <ell Eabs6 he had the -roblem of analy8in# the 9$,F 5ernel
with res-ect to distributed com-utin#! "ememberin# his 3h!)! e0-erience6 Stroustru- set
out to enhance the C lan#ua#e with Simula>li5e features! C was chosen because it was
#eneral>-ur-ose6 fast6 -ortable and widely used! <esides C and Simula6 some other
lan#ua#es that ins-ired him were .E1E C(6 .da6 CE9 and :E! .t first6 the class6
derived class6 stron# ty-e chec5in#6 inlinin#6 and default ar#ument features were added to
C via Cfront! +he first commercial release occurred in 1ctober 1=(;!
,n 1=(36 the name of the lan#ua#e was chan#ed from C with Classes to C?? @?? bein#
the increment o-erator in C and C??A! $ew features were added includin# virtual
functions6 function name and o-erator overloadin#6 references6 constants6 user>controlled
free>store memory control6 im-roved ty-e chec5in#6 and <C3E style sin#le>line
comments with two forward slashes @&&A! ,n 1=(;6 the first edition of +he C??
3ro#rammin# Ean#ua#e was released6 -rovidin# an im-ortant reference to the lan#ua#e6
since there was not yet an official standard! "elease 2!0 of C?? came in 1=(=! $ew
features included multi-le inheritance6 abstract classes6 static member functions6 const
member functions6 and -rotected members! ,n 1==06 +he .nnotated C?? "eference
:anual was -ublished! +his wor5 became the basis for the future standard! Eate addition
of features included tem-lates6 e0ce-tions6 names-aces6 new casts6 and a <oolean ty-e!
.s the C?? lan#ua#e evolved6 a standard library also evolved with it! +he first addition to
the C?? standard library was the stream ,&1 library which -rovided facilities to re-lace
the traditional C functions such as -rintf and scanf! Eater6 amon# the most si#nificant
additions to the standard library6 was the Standard +em-late Eibrary!
1=
C?? continues to be used and is one of the -referred -ro#rammin# lan#ua#es to develo-
-rofessional a--lications! +he lan#ua#e has #one from bein# mostly Western to attractin#
-ro#rammers from all over the world!
Language standard
,n 1==(6 the C?? standards committee @the ,S1&,EC J+C1&SC22&W21 wor5in# #rou-A
standardi8ed C?? and -ublished the international standard ,S1&,EC 1'((221==(
@informally 5nown as C??=(V'WA! 7or some years after the official release of the standard6
the committee -rocessed defect re-orts6 and -ublished a corrected version of the C??
standard6 ,S1&,EC 1'((2220036 in 2003! ,n 200;6 a technical re-ort6 called the GEibrary
+echnical "e-ort 1G @often 5nown as +"1 for shortA6 was released! While not an official
-art of the standard6 it s-ecified a number of e0tensions to the standard library6 which
were e0-ected to be included in the ne0t version of C??! Su--ort for +"1 is #rowin# in
almost all currently maintained C?? com-ilers!
While the C?? lan#ua#e is royalty>free6 the standard document itself is not freely
available!
Etymology
.ccordin# to Stroustru-2 Gthe name si#nifies the evolutionary nature of the chan#es from
CG! )urin# C??Rs develo-ment -eriod6 the lan#ua#e had been referred to as Gnew CG6
then GC with ClassesG! +he final name is credited to "ic5 :ascitti @mid>1=(3A and was
first used in )ecember 1=(3! When :ascitti was 4uestioned informally in 1==2 about the
namin#6 he indicated that it was #iven in a ton#ue>in>chee5 s-irit! ,t stems from CRs G??G
o-erator @which increments the value of a variableA and a common namin# convention of
usin# G?G to indicate an enhanced com-uter -ro#ram! +here is no lan#ua#e called GC
-lusG! .<CE&c? was the name of an earlier6 unrelated -ro#rammin# lan#ua#e!
20
!ilosop!y
,n +he )esi#n and Evolution of C?? @1=='A6 </arne Stroustru- describes some rules that
he uses for the desi#n of C??2
C?? is desi#ned to be a statically ty-ed6 #eneral>-ur-ose lan#ua#e that is as efficient and
-ortable as C!

C?? is desi#ned to directly and com-rehensively su--ort multi-le -ro#rammin# styles
@-rocedural -ro#rammin#6 data abstraction6 ob/ect>oriented -ro#rammin#6 and #eneric
-ro#rammin#A
C?? is desi#ned to #ive the -ro#rammer choice6 even if this ma5es it -ossible for the
-ro#rammer to choose incorrectly
C?? is desi#ned to be as com-atible with C as -ossible6 therefore -rovidin# a smooth
transition from C
C?? avoids features that are -latform s-ecific or not #eneral -ur-ose
C?? does not incur overhead for features that are not used @the G8ero>overhead
-rinci-leGA
C?? is desi#ned to function without a so-histicated -ro#rammin# environment
,nside the C?? 1b/ect :odel @Ei--man6 1==CA describes how com-ilers may convert C?
? -ro#ram statements into an in>memory layout! Com-iler authors are6 however6 free to
im-lement the standard in their own manner!
Standard library
+he 1==( .$S,&,S1 C?? standard consists of two -arts2 the core lan#ua#e and the C??
Standard EibraryN the latter includes most of the Standard +em-late Eibrary @S+EA and a
21
sli#htly modified version of the C standard library! :any C?? libraries e0ist which are
not -art of the standard6 and6 usin# lin5a#e s-ecification6 libraries can even be written in
lan#ua#es such as C6 7ortran6 3ascal6 or <.S,C! Which of these are su--orted is
com-iler de-endent!
+he C?? standard library incor-orates the C standard library with some small
modifications to ma5e it o-timi8ed with the C?? lan#ua#e! .nother lar#e -art of the C??
library is based on the S+E! +his -rovides such useful tools as containers @for e0am-le
vectors and listsA6 iterators to -rovide these containers with array>li5e access and
al#orithms to -erform o-erations such as searchin# and sortin#! 7urthermore @multiAma-s
@associative arraysA and @multiAsets are -rovided6 all of which e0-ort com-atible
interfaces! +herefore it is -ossible6 usin# tem-lates6 to write #eneric al#orithms that wor5
with any container or on any se4uence defined by iterators! .s in C6 the features of the
library are accessed by usin# the Binclude directive to include a standard header! C??
-rovides C= standard headers6 of which 1= are de-recated!
+he S+E was ori#inally a third>-arty library from D3 and later S,6 before its
incor-oration into the C?? standard! +he standard does not refer to it as GS+EG6 as it is
merely a -art of the standard library6 but many -eo-le still use that term to distin#uish it
from the rest of the library @in-ut&out-ut streams6 internationali8ation6 dia#nostics6 the C
library subset6 etc!A!
:ost C?? com-ilers -rovide an im-lementation of the C?? standard library6 includin#
the S+E! Com-iler>inde-endent im-lementations of the S+E6 such as S+E3ort6 also e0ist!
1ther -ro/ects also -roduce various custom im-lementations of the C?? standard library
and the S+E with various desi#n #oals!
Language features
C?? inherits most of CRs synta0 and the C -re-rocessor! +he followin# is </arne
Stroustru-Rs versionRs of the Dello world -ro#ram which uses the C?? standard library
stream facility to write a messa#e to standard out-ut2
22
Binclude PiostreamQ

int main@A
X
std22cout PP GDello6 worldJSnGN
Y
"perators and operator o#erloading
C?? -rovides more than 30 o-erators6 coverin# basic arithmetic6 bit mani-ulation6
indirection6 com-arisons6 lo#ical o-erations and more! .lmost all o-erators can be
overloaded for user>defined ty-es6 with a few notable e0ce-tions such as member access
@! and !OA! +he rich set of overloadable o-erators is central to usin# C?? as a domain
s-ecific lan#ua#e! +he overloadable o-erators are also an essential -art of many advanced
C?? -ro#rammin# techni4ues6 such as smart -ointers! 1verloadin# an o-erator does not
chan#e the -recedence of calculations involvin# the o-erator6 nor does it chan#e the
number of o-erands that the o-erator uses @any o-erand may however be i#noredA!
Templates
C?? tem-lates enable #eneric -ro#rammin#! C?? su--orts both function and class
tem-lates! +em-lates may be -arameteri8ed by ty-es6 com-ile>time constants6 and other
tem-lates! C?? tem-lates are im-lemented by instantiation at com-ile>time! +o
instantiate a tem-late6 com-ilers substitute s-ecific ar#uments for a tem-lateRs -arameters
to #enerate a concrete function or class instance! +em-lates are a -owerful tool that can
be used for #eneric -ro#rammin#6 tem-late meta-ro#rammin#6 and code o-timi8ation6
but this -ower im-lies a cost! +em-late use may increase code si8e6 since each tem-late
instantiation -roduces a co-y of the tem-late code2 one for each set of tem-late
ar#uments! +his is in contrast to run>time #enerics seen in other lan#ua#es @e!#! JavaA
where at com-ile>time the ty-e is erased and a sin#le tem-late body is -reserved!
23
+em-lates are different from macros2 while both of these com-ile>time lan#ua#e features
enable conditional com-ilation6 tem-lates are not restricted to le0ical substitution!
+em-lates are aware of the semantics and ty-e system of their com-anion lan#ua#e6 as
well as all com-ile>time ty-e definitions6 and can -erform hi#h>level o-erations includin#
-ro#rammatic flow control based on evaluation of strictly ty-e>chec5ed -arameters!
:acros are ca-able of conditional control over com-ilation based on -redetermined
criteria6 but cannot instantiate new ty-es6 recurse6 or -erform ty-e evaluation and in effect
are limited to -re>com-ilation te0t>substitution and te0t>inclusion&e0clusion! ,n other
words6 macros can control com-ilation flow based on -re>defined symbols but cannot6
unli5e tem-lates6 inde-endently instantiate new symbols! +em-lates are a tool for static
-olymor-hism @see belowA and #eneric -ro#rammin#!
,n addition6 tem-lates are a com-ile time mechanism in C?? which is +urin#>com-lete6
meanin# that any com-utation e0-ressible by a com-uter -ro#ram can be com-uted6 in
some form6 by a tem-late meta-ro#ram -rior to runtime!
,n summary6 a tem-late is a com-ile>time -arameteri8ed function or class written without
5nowled#e of the s-ecific ar#uments used to instantiate it! .fter instantiation the resultin#
code is e4uivalent to code written s-ecifically for the -assed ar#uments! ,n this manner6
tem-lates -rovide a way to decou-le #eneric6 broadly>a--licable as-ects of functions and
classes @encoded in tem-latesA from s-ecific as-ects @encoded in tem-late -arametersA
without sacrificin# -erformance due to abstraction!
"b$ects
C?? introduces ob/ect>oriented @11A features to C! ,t offers classes6 which -rovide the
four features commonly -resent in 11 @and some non>11A lan#ua#es2 abstraction6
enca-sulation6 inheritance6 and -olymor-hism! 1b/ects are instances of classes created at
runtime! +he class can be thou#ht of as a tem-late from which many different individual
ob/ects may be #enerated as a -ro#ram runs!
2'
Encapsulation
Enca-sulation is the hidin# of information in order to ensure that data structures and
o-erators are used as intended and to ma5e the usa#e model more obvious to the
develo-er! C?? -rovides the ability to define classes and functions as its -rimary
enca-sulation mechanisms! Within a class6 members can be declared as either -ublic6
-rotected6 or -rivate in order to e0-licitly enforce enca-sulation! . -ublic member of the
class is accessible to any function! . -rivate member is accessible only to functions that
are members of that class and to functions and classes e0-licitly #ranted access
-ermission by the class @GfriendsGA! . -rotected member is accessible to members of
classes that inherit from the class in addition to the class itself and any friends!
+he 11 -rinci-le is that all of the functions @and only the functionsA that access the
internal re-resentation of a ty-e should be enca-sulated within the ty-e definition! C??
su--orts this @via member functions and friend functionsA6 but does not enforce it2 the
-ro#rammer can declare -arts or all of the re-resentation of a ty-e to be -ublic6 and is
allowed to ma5e -ublic entities that are not -art of the re-resentation of the ty-e! <ecause
of this6 C?? su--orts not /ust 11 -ro#rammin#6 but other wea5er decom-osition
-aradi#ms6 li5e modular -ro#rammin#!
,t is #enerally considered #ood -ractice to ma5e all data -rivate or -rotected6 and to ma5e
-ublic only those functions that are -art of a minimal interface for users of the class! +his
hides all the details of data im-lementation6 allowin# the desi#ner to later fundamentally
chan#e the im-lementation without chan#in# the interface in any way!V(WV=W
%n!eritance
,nheritance allows one data ty-e to ac4uire -ro-erties of other data ty-es! ,nheritance
from a base class may be declared as -ublic6 -rotected6 or -rivate! +his access s-ecifier
determines whether unrelated and derived classes can access the inherited -ublic and
-rotected members of the base class! 1nly -ublic inheritance corres-onds to what is
usually meant by GinheritanceG! +he other two forms are much less fre4uently used! ,f the
access s-ecifier is omitted6 a GclassG inherits -rivately6 while a GstructG inherits -ublicly!
2;
<ase classes may be declared as virtualN this is called virtual inheritance! %irtual
inheritance ensures that only one instance of a base class e0ists in the inheritance #ra-h6
avoidin# some of the ambi#uity -roblems of multi-le inheritance!
:ulti-le inheritance is a C?? feature sometimes considered controversial! :ulti-le
inheritance allows a class to be derived from more than one base classN this can result in a
com-licated #ra-h of inheritance relationshi-s! 7or e0am-le6 a G7lyin# CatG class can
inherit from both GCatG and G7lyin# :ammalG! Some other lan#ua#es6 such as Java or
CB6 accom-lish somethin# similar @althou#h more limitedA by allowin# inheritance of
multi-le interfaces while restrictin# the number of base classes to one @interfaces6 unli5e
classes6 -rovide only declarations of member functions6 no im-lementation or member
dataA! ,nterfaces and abstract classes in Java and CB can be e0-ressed in C?? as a class
containin# only function declarations6 often 5nown as an abstract base class or G.<C!G
3ro#rammers -referrin# the Java&CB model of inheritance can choose to inherit only one
non>abstract class6 and can otherwise limit themselves to only abstract classes!
olymorp!ism
See also2 3olymor-hism in ob/ect>oriented -ro#rammin#
3olymor-hism enables one common interface for many im-lementations6 and for ob/ects
to act differently under different circumstances!
C?? su--orts several 5inds of static @com-ile>timeA and dynamic @run>timeA
-olymor-hisms! Com-ile>time -olymor-hism does not allow for certain run>time
decisions6 while run>time -olymor-hism ty-ically incurs a -erformance -enalty!
Static polymorp!ism
7unction overloadin# allows -ro#rams to declare multi-le functions havin# the same
name @but with different ar#umentsA! +he functions are distin#uished by the number
and&or ty-es of their formal -arameters! +hus6 the same function name can refer to
different functions de-endin# on the conte0t in which it is used! +he ty-e returned by the
function is not used to distin#uish overloaded functions!
2C
When declarin# a function6 a -ro#rammer can s-ecify default ar#uments for one or more
-arameters! )oin# so allows the -arameters with defaults to o-tionally be omitted when
the function is called6 in which case the default ar#uments will be used! When a function
is called with fewer ar#uments than there are declared -arameters6 e0-licit ar#uments are
matched to -arameters in left>to>ri#ht order6 with any unmatched -arameters at the end of
the -arameter list bein# assi#ned their default ar#uments! ,n many cases6 s-ecifyin#
default ar#uments in a sin#le function declaration is -referable to -rovidin# overloaded
function definitions with different numbers of -arameters!
+em-lates in C?? -rovide a so-histicated mechanism for writin# #eneric6 -olymor-hic
code! ,n -articular6 throu#h the Curiously "ecurrin# +em-late 3attern itRs -ossible to
im-lement a form of static -olymor-hism that closely mimics the synta0 for overridin#
virtual functions! Since C?? tem-lates are ty-e>aware and +urin#>com-lete they can also
be used to let the com-iler resolve recursive conditionals and #enerate substantial
-ro#rams throu#h tem-late meta-ro#rammin#!
&ynamic polymorp!ism
%n!eritance
%ariable -ointers @and referencesA to a base class ty-e in C?? can refer to ob/ects of any
derived classes of that ty-e in addition to ob/ects e0actly matchin# the variable ty-e! +his
allows arrays and other 5inds of containers to hold -ointers to ob/ects of differin# ty-es!
<ecause assi#nment of values to variables usually occurs at run>time6 this is necessarily a
run>time -henomenon!
C?? also -rovides a dynamic\cast o-erator6 which allows the -ro#ram to safely attem-t
conversion of an ob/ect into an ob/ect of a more s-ecific ob/ect ty-e @as o--osed to
conversion to a more #eneral ty-e6 which is always allowedA! +his feature relies on run>
time ty-e information @"++,A! 1b/ects 5nown to be of a certain s-ecific ty-e can also be
cast to that ty-e with static\cast6 a -urely com-ile>time construct which is faster and does
not re4uire "++,!
2*
'irtual member functions
1rdinarily when a function in a derived class overrides a function in a base class6 the
function to call is determined by the ty-e of the ob/ect! . #iven function is overridden
when there e0ists no difference6 in the number or ty-e of -arameters6 between two or
more definitions of that function! Dence6 at com-ile time it may not be -ossible to
determine the ty-e of the ob/ect and therefore the correct function to call6 #iven only a
base class -ointerN the decision is therefore -ut off until runtime! +his is called dynamic
dis-atch! %irtual member functions or methods allow the most s-ecific im-lementation of
the function to be called6 accordin# to the actual run>time ty-e of the ob/ect! ,n C??6 this
is commonly done usin# virtual function tables! ,f the ob/ect ty-e is 5nown6 this may be
by-assed by -re-endin# a fully 4ualified class name before the function call6 but in
#eneral calls to virtual functions are resolved at run time!
,n addition to standard member functions6 o-erator overloads and destructors can be
virtual! . #eneral rule of thumb is that if any functions in the class are virtual6 the
destructor should be as well! .s the ty-e of an ob/ect at its creation is 5nown at com-ile
time6 constructors6 and by e0tension co-y constructors6 cannot be virtual! $onetheless a
situation may arise where a co-y of an ob/ect needs to be created when a -ointer to a
derived ob/ect is -assed as a -ointer to a base ob/ect! ,n such a case a common solution is
to create a clone@A @or similarA function and declare that as virtual! +he clone@A method
creates and returns a co-y of the derived class when called!
. member function can also be made G-ure virtualG by a--endin# it with M 0 after the
closin# -arenthesis and before the semicolon! 1b/ects cannot be created of a class with a
-ure virtual function and are called abstract data ty-es! Such abstract data ty-es can only
be derived from! .ny derived class inherits the virtual function as -ure and must -rovide
a non>-ure definition of it @and all other -ure virtual functionsA before ob/ects of the
derived class can be created! .n attem-t to create an ob/ect from a class with a -ure
virtual function or inherited -ure virtual function will be fla##ed as a com-ile>time error!
2(
arsing and processing C(( source code
,t is relatively difficult to write a #ood C?? -arser with classic -arsin# al#orithms such as
E.E"@1A! +his is -artly because the C?? #rammar is not E.E"! <ecause of this6 there
are very few tools for analy8in# or -erformin# non>trivial transformations @e!#!6
refactorin#A of e0istin# code! 1ne way to handle this difficulty is to choose a different
synta06 such as Si#nificantly 3rettier and Easier C?? Synta06 which is E.E"@1A -arsable!
:ore -owerful -arsers6 such as E" -arsers6 can be substantially sim-ler @thou#h
slowerA!
3arsin# @in the literal sense of -roducin# a synta0 treeA is not the most difficult -roblem
in buildin# a C?? -rocessin# tool! Such tools must also have the same understandin# of
the meanin# of the identifiers in the -ro#ram as a com-iler mi#ht have! 3ractical systems
for -rocessin# C?? must then not only -arse the source te0t6 but be able to resolve for
each identifier -recisely which definition a--lies @e!#! they must correctly handle C??Rs
com-le0 sco-in# rulesA and what its ty-e is6 as well as the ty-es of lar#er e0-ressions!
7inally6 a -ractical C?? -rocessin# tool must be able to handle the variety of C?? dialects
used in -ractice @such as that su--orted by the $9 Com-iler Collection and that of
:icrosoftRs %isual C??A and im-lement a--ro-riate analy8ers6 source code transformers6
and re#enerate source te0t! Combinin# advanced -arsin# al#orithms such as E" with
symbol table construction and -ro#ram transformation machinery can enable the
construction of arbitrary C?? tools!
Compatibility
3roducin# a reasonably standards>com-liant C?? com-iler has -roven to be a difficult
tas5 for com-iler vendors in #eneral! 7or many years6 different C?? com-ilers
im-lemented the C?? lan#ua#e to different levels of com-liance to the standard6 and their
im-lementations varied widely in some areas such as -artial tem-late s-eciali8ation!
"ecent releases of most -o-ular C?? com-ilers su--ort almost all of the C?? 1==(
standard! .--arently there is no im-lementation that fully su--orts the entire standard!
2=
1ne -articular -oint of contention is the e0-ort 5eyword6 intended to allow tem-late
definitions to be se-arated from their declarations! +he first com-iler to im-lement e0-ort
was Comeau C&C??6 in early 2003 @; years after the release of the standardAN in 200'6 the
beta com-iler of <orland C?? <uilder F was also released with e0-ort! <oth of these
com-ilers are based on the E) C?? front end! ,t should also be noted that many C??
boo5s -rovide e0am-le code usin# the 5eyword e0-ort @for e0am-le6 <e#innin# .$S, C?
? by ,vor DortonA which will not com-ile in most com-ilers6 but there is no reference to
the -roblem with the 5eyword e0-ort mentioned! 1ther com-ilers such as CC do not
su--ort it at all! Derb Sutter6 former convener of the C?? standards committee6
recommended that e0-ort be removed from future versions of the C?? standard6 but
finally the decision was made to retain it!
,n order to #ive com-iler vendors #reater freedom6 the C?? standards committee decided
not to dictate the im-lementation of name man#lin#6 e0ce-tion handlin#6 and other
im-lementation>s-ecific features! +he downside of this decision is that ob/ect code
-roduced by different com-ilers is e0-ected to be incom-atible! +here are6 however6 third
-arty standards for -articular machines or o-eratin# systems which attem-t to standardi8e
com-ilers on those -latforms @for e0am-le C?? .<,V1;WAN some com-ilers ado-t a
secondary standard for these items!
)it! C
7or more details on this to-ic6 see Com-atibility of C and C??!
C?? is often considered to be a su-erset of C6 but this is not strictly true!V1CW :ost C
code can easily be made to com-ile correctly in C??6 but there are a few differences that
cause some valid C code to be invalid in C??6 or to behave differently in C??!
1ne commonly encountered difference is that C allows im-licit conversion from voidO to
other -ointer ty-es6 but C?? does not! .nother common -ortability issue is that C??
defines many new 5eywords6 such as new and class6 that may be used as identifiers @e!#!
variable namesA in a C -ro#ram!
30
Some incom-atibilities have been removed by the latest @C==A C standard6 which now
su--orts C?? features such as && comments and mi0ed declarations and code! 1n the
other hand6 C== introduced a number of new features that C?? does not su--ort6 such as
variable>len#th arrays6 native com-le0>number ty-es6 desi#nated initiali8ers and
com-ound literals! Dowever6 at least some of the new C== features will li5ely be
included in the ne0t version of the C?? standard6 C??00!
,n order to intermi0 C and C?? code6 any function declaration or definition that is to be
called from&used both in C and C?? must be declared with C lin5a#e by -lacin# it within
an e0tern GCG X !!! Y bloc5! Such a function may not rely on features de-endin# on name
man#lin# @i!e!6 function overloadin#A!
31
.. PRO*RAMMIN* IN C++
%*T+"&,CT%"* T" C((
C?? is lan#ua#e develo-ed to su--ort modern 113Is6 while at the same time
retainin# the loo5 of already e0istin# and 4uite -o-ular lan#ua#e ZCI! +his section
-rovides an overview of the 5ey conce-ts embodied in C??! ,ts ob/ect oriented features
are hi#hly interrelated! 1n several instances one feature im-lies -rior 5nowled#e of one
or more features! +his section #ives a 4uic5 overview of most im-ortant as-ects of C??!
"+%-%* "F C((
C?? is an e0-anded version of C! +he C?? e0tension to C were first invented by
<.J."$E S+"19S+"93 in 1=(0 at <ell labs !.lthou#h C ??Is -redecessor is one of
most li5ed and widely used -rofessional -ro#rammin# factors increasin# com-le0ity!
1ver the years com-uters -ro#rams has become lar#e and com-le0!ity !essence of C?? is
to allow -ro#rammer to com-rehend and mana#e lar#er and com-le0 -ro#ram!
When C?? was invented6 STROUSTRUP 5new that it was im-ortant to maintain
ori#inal s-rit of C6 includin# its efficiency6 its fle0ibility and under lyin# -hiloso-hy that
the -ro#rams6 not the lan#ua#e6 is in chan#e while at the same time addin# su--ort for
113Is! Da--ily6 his #oals were accom-lished !C?? still -rovides the -ro#rammer with
the freedom and control of C?? still -rovides the -ro#rammer with the freedom and
control of C cou-led with the -ower of ob/ect!
)./T %S ""S0

1b/ect 1riented 3ro#rammin# @113A is a new way of a--roachin# the /ob of the
-ro#rams! .--roaches to -ro#rammin# have chan#ed dramatically since the invention of
com-uter6 -rimarily to accommodate the increasin# com-le0ity of -ro#rams! 7or
e0am-le 6the -ro#rammin# a--roach evolved from the machine lan#ua#e -ro#rammin#
involves DEEs @hi#h level lan#ua#e A! Certain hi#hli#hts of structured -ro#rammin#
are2
32
1A em-hasis on doin# thin#s rather than data!
2A Ear#e -ro#ram divided into smaller -ro#rams 5nown as functions!
3A )ata is shared fre4uently !
'A )ata moves o-enly around the system from function to function!
;A 7unction can transform data one from to another!
113 has ta5en the best ideas of structured -ro#rammin# and combined with several
-owerful conce-ts that encoura#es you to a--roach the tas5 of -ro#rammin# in a
new way!
When -ro#rammin# in a ob/ect oriented manner 6a -roblem is bro5en down into
sub#rou-s of the related -arts that ta5e into account both code and data related to
each #rou-! .lso these sub#rou-s are or#ani8ed into a hierarchical structure! 7inally
these sub#rou-s are translated into self>contained units called ob/ects!
.ll ob/ects oriented -ro#rammin# lan#ua#es have three traits in common2
Enca-sulation6 -olymor-hism and inheritance ! Eet us e0amine these briefly!
C++ at a *"ance
Some of im-ortant features or im-ortant -arts of C?? !
1A 9nli5e C6C?? is a stron#ly ty-ed lan#ua#e 2 Every ob/ect must belon# to a ty-e
and an ob/ect of one ty-e cannot be assi#ned to or com-ared with one of the
another ty-es by well defined conversion!
2A 7unction -rototy-e is mandatory in C??!
3A Call by reference is su--orted in C??! +his removes one of the difficult -ointer
code sometimes found in C code ! . direct result is that any lar#e structure should
be -assed by reference!
'A +he com-iler directive const can be used in many -laces to im-ly that some
ob/ects is unalterable ! +his is manifested in no ! of ways! 7irst lar#e structure can
be safely -assed as const and by reference 6 rather than usin# a -ointer ! Second
33
even inte#ers that are -assed by value can be s-ecified as const to -revent
accidental modification!
.-art from the above characteristics6 s-ecific constructs that are offered by C?? are
discussed later on!
Classes
. class in C?? is collection of some data fields and the functions mani-ulatin# them ! .
class may contain -ublic as well -rivate -arts !<y default6 all the items defined in a class
are -rivate !+his means that they cannot be accessed by any function that is not a member
of the class !+his is one way that enca-sulation has been achieved ] access to certain
items of data can be ti#htly controlled by 5ee-in# them -rivate!
+o ma5es -arts of a class -ublic 6 these must be declared after the -ublic 5eyword !.ll
variables or function defined after -ublic can be accessed by other function in the
-ro#ram ! Essentially 6 the rest of your -ro#ram access an ob/ect throu#h its -ublic
function!
"'E+L"/&%*-
C?? achieves -olymor-hism throu#h the use of overloadin# ! ,n C?? overloadin# is of
two ty-es 2 7unctional overloadin# and o-erator overloadin#!
,t is allowed in C?? that two or more function share name as lon# as their -arameter
declaration are different ! ,n this situation the function that shares the same name said to
be overloaded and this -rocess is referred to as function overloadin# ! When a function
call is made the com-iler determines which of the identical names are bein# referred to!

Closely related to functional overloadin# there is o-erator overloadin# !,n C??6 you can
overload most o-erators so that they can -erform s-ecial o-erations relative to classes
3'
that you create! 7or e0am-le a class that maintains a stac5 mi#ht overload ?? to -erform
-ush u- o-eration and ] to -erform a -o-! When an o-erator is overloaded none of its
ori#inal meanin# are lost ! ,nstead the ty-e of ob/ects it can be a--lied to is e0-anded!
1-erator overloadin# is achieved by creatin# o-erator function! .n o-erator function
defines o-erations that the overloaded o-erator will -erform relative to the class it is
desi#ned to wor5 on!
C"*ST+,CT%"+S /*& &EST+,CT%"*S
,t is very common form some -art of an ob/ect to re4uire initiali8ation before it
can be used ! <ecause of the re4uirements for initiali8ation is so common 6C??
allows ob/ects to initiali8e themselves when they are created !+he automatic
initiali8ation is -erformed throu#h the use of constructor function !. constructor
function is a s-ecial function that is a member of a class and has the same name as
that of the class !.s the ob/ectI s constructor is automatically called when the
ob/ect is created ! +his means that it is called when the ob/ectIs declaration is
e0ecuted!
+he com-lement of the constructor is destructor ! ,n many circumstances
an ob/ect will need to -erform some action or actions when it is destroyed !7or
e0am-le closin# of an o-ened file deallocation of memory etc ! ,n C?? it is the
destructor function that handles deactivation events! +he destructor has the same
name as that of the constructor6 but it is -receded by> @tildeA symbol!
+"-+/11%*-
3ro#rammin# is the course of desi#nin# 6writin#6 testin# and maintainin# that
ma5es the com-uter wor5s accordin# to our re4uirements! +his section e0-lains
how efficient -ro#rammin# can be achieved!
3;
W%at i# a Progra./

1A . -lan to achieve a -roblem solution N
2A +o desi#n6 write and test one or more routines N
3A . set of se4uenced instructions to cause a com-uter to -erform -articular
o-eration!
,o0 to 0rite a C++ Progra /
3ro#rams are not 4uic5 creations ! "ather a series of ste-s that are to be wor5ed
out 6 which are bein# ta5en by the command lines!
Ste! 1 .De1ine t%e need 2> +o create a new -ro#ram or to im-rove or to modify an
e0istin# one 6 first reco#ni8e that -roblem ! 9nder this ste- the -ro#rammer must have a
clear understandin# of e0actly what is that to be done T
Ste! 2. Ana"23e 2> +his ste- is basically concerned with the decom-osition of a
-roblem into the com-onents -arts and establishin# a lo#ical model to describe the
system functions ! 7ollowin# are the sub>ste-s constitute this ste- !

1A ,dentifyin# the ob/ects and their attributes!
2A ,dentifyin# the service that each ob/ect is e0-ected to -rovide!
3A Establishin# inter ]connection between ob/ects in terms of service re4uired
and services rendered!
Ste! $.De#ign 2> 9nder this ste- followin# actions ta5e -lace2

aA S-ecifications of the class de-endencies!
bA 1r#ani8ation of the class hierarchies!
cA )esi#n of classes !
dA )esi#n of member functions !
eA )esi#n of the driver -ro#ram !
3C
Every C?? -ro#ram must ain45 function code 5nown as dri6e !rogra ! +he
e0ecution of the -ro#ram be#ins and ends here! +he drive -ro#ram is mainly res-onsible
for N

aA "eceivin# data values from the userN
bA Creatin# ob/ects from class definitionsN
cA Establishin# communication between ob/ectsN
dA )is-layin# & -roducin# the out-ut results in the form re4uired by the user! ,n
nutshell6 the driver -ro#ram is #ateway to the users!
Ste! ). I!"eent :7+his includes codin# and testin# ! Codin# includes writin#
codes for classes and member function and the ain -ro#ram which acts as a driver
in the -ro#ram ! +o develo- the lo#ic of the functions6 -ro#ram develo-ment tools
may be used!
$o -ro#ram wor5s accurately for the first time ! So testin# the -ro#ram
before usin# is an essential -art of the -ro#rammin# !+he final #oal of testin# is to see
that the -ro#ram the intended /ob satisfactorily!

3*
8. Str&ct&re o1 a !rogra
3robably the best way to start learnin# a -ro#rammin# lan#ua#e is by writin# a -ro#ram!
+herefore6 here is our first -ro#ram2
&& my first -ro#ram in C??
Binclude PiostreamQ
usin# names-ace stdN
int main @A
X
cout PP GDello WorldJGN
return 0N
Y
Dello WorldJ
+he first -anel shows the source code for our first -ro#ram! +he second one shows the
result of the -ro#ram once com-iled and e0ecuted! +he way to edit and com-ile a
-ro#ram de-ends on the com-iler you are usin#! )e-endin# on whether it has a
)evelo-ment ,nterface or not and on its version! Consult the com-ilers section and the
manual or hel- included with your com-iler if you have doubts on how to com-ile a C??
console -ro#ram!
+he -revious -ro#ram is the ty-ical -ro#ram that -ro#rammer a--rentices write for the
first time6 and its result is the -rintin# on screen of the GDello WorldJG sentence! ,t is one
of the sim-lest -ro#rams that can be written in C??6 but it already contains the
fundamental com-onents that every C?? -ro#ram has! We are #oin# to loo5 line by line
at the code we have /ust written2
99 2 1ir#t !rogra in C++
3(
+his is a comment line! .ll lines be#innin# with two slash si#ns @&&A are
considered comments and do not have any effect on the behavior of the -ro#ram!
+he -ro#rammer can use them to include short e0-lanations or observations
within the source code itself! ,n this case6 the line is a brief descri-tion of what
our -ro#ram is!
:inc"&de ;io#trea<
Eines be#innin# with a hash si#n @BA are directives for the -re-rocessor! +hey are
not re#ular code lines with e0-ressions but indications for the com-ilerRs
-re-rocessor! ,n this case the directive Binclude PiostreamQ tells the -re-rocessor
to include the iostream standard file! +his s-ecific file @iostreamA includes the
declarations of the basic standard in-ut>out-ut library in C??6 and it is included
because its functionality is #oin# to be used later in the -ro#ram!
&#ing nae#!ace #td=
.ll the elements of the standard C?? library are declared within what is called a
names-ace6 the names-ace with the name std! So in order to access its
functionality we declare with this e0-ression that we will be usin# these entities!
+his line is very fre4uent in C?? -ro#rams that use the standard library6 and in
fact it will be included in most of the source codes included in these tutorials!
int ain 45
+his line corres-onds to the be#innin# of the definition of the main function! +he
main function is the -oint by where all C?? -ro#rams start their e0ecution6
inde-endently of its location within the source code! ,t does not matter whether
there are other functions with other names defined before or after it > the
instructions contained within this functionRs definition will always be the first
ones to be e0ecuted in any C?? -ro#ram! 7or that same reason6 it is essential that
all C?? -ro#rams have a main function!
+he word main is followed in the code by a -air of -arentheses @@AA! +hat is
because it is a function declaration2 ,n C??6 what differentiates a function
3=
declaration from other ty-es of e0-ressions are these -arentheses that follow its
name! 1-tionally6 these -arentheses may enclose a list of -arameters within them!
"i#ht after these -arentheses we can find the body of the main function enclosed
in braces @XYA! What is contained within these braces is what the function does
when it is e0ecuted!
co&t ;; >,e""o Wor"d>=
+his line is a C?? statement! . statement is a sim-le or com-ound e0-ression that
can actually -roduce some effect! ,n fact6 this statement -erforms the only action
that #enerates a visible effect in our first -ro#ram!
cout re-resents the standard out-ut stream in C??6 and the meanin# of the entire
statement is to insert a se4uence of characters @in this case the Dello World
se4uence of charactersA into the standard out-ut stream @which usually is the
screenA!
cout is declared in the iostream standard file within the std names-ace6 so thatRs
why we needed to include that s-ecific file and to declare that we were #oin# to
use this s-ecific names-ace earlier in our code!
$otice that the statement ends with a semicolon character @NA! +his character is
used to mar5 the end of the statement and in fact it must be included at the end of
all e0-ression statements in all C?? -ro#rams @one of the most common synta0
errors is indeed to for#et to include some semicolon after a statementA!
?aria@"e#. Data T2!e#.
+he usefulness of the GDello WorldG -ro#rams shown in the -revious section is 4uite
4uestionable! We had to write several lines of code6 com-ile them6 and then e0ecute the
resultin# -ro#ram /ust to obtain a sim-le sentence written on the screen as result! ,t
certainly would have been much faster to ty-e the out-ut sentence by ourselves!
Dowever6 -ro#rammin# is not limited only to -rintin# sim-le te0ts on the screen! ,n order
'0
to #o a little further on and to become able to write -ro#rams that -erform useful tas5s
that really save us wor5 we need to introduce the conce-t of variable!
Eet us thin5 that , as5 you to retain the number ; in your mental memory6 and then , as5
you to memori8e also the number 2 at the same time! Hou have /ust stored two different
values in your memory! $ow6 if , as5 you to add 1 to the first number , said6 you should
be retainin# the numbers C @that is ;?1A and 2 in your memory! %alues that we could now
for e0am-le subtract and obtain ' as result!
+he whole -rocess that you have /ust done with your mental memory is a simile of what a
com-uter can do with two variables! +he same -rocess can be e0-ressed in C?? with the
followin# instruction set2
a M ;N
b M 2N
a M a ? 1N
result M a > bN
1bviously6 this is a very sim-le e0am-le since we have only used two small inte#er
values6 but consider that your com-uter can store millions of numbers li5e these at the
same time and conduct so-histicated mathematical o-erations with them!
+herefore6 we can define a variable as a -ortion of memory to store a determined value!
Each variable needs an identifier that distin#uishes it from the others6 for e0am-le6 in the
-revious code the variable identifiers were a6 b and result6 but we could have called the
variables any names we wanted to invent6 as lon# as they were valid identifiers!
Dec"aration o1 6aria@"e#
,n order to use a variable in C??6 we must first declare it s-ecifyin# which data ty-e we
want it to be! +he synta0 to declare a new variable is to write the s-ecifier of the desired
data ty-e @li5e int6 bool6 float!!!A followed by a valid variable identifier! 7or e0am-le2
'1
int aN
float mynumberN
+hese are two valid declarations of variables! +he first one declares a variable of ty-e int
with the identifier a! +he second one declares a variable of ty-e float with the identifier
mynumber! 1nce declared6 the variables a and mynumber can be used within the rest of
their sco-e in the -ro#ram!
,f you are #oin# to declare more than one variable of the same ty-e6 you can declare all of
them in a sin#le statement by se-aratin# their identifiers with commas! 7or e0am-le2
int a6 b6 cN
+his declares three variables @a6 b and cA6 all of them of ty-e int6 and has e0actly the same
meanin# as2
int aN
int bN
int cN
Introd&ction to #tring#
%ariables that can store non>numerical values that are lon#er than one sin#le character are
5nown as strin#s!
+he C?? lan#ua#e library -rovides su--ort for strin#s throu#h the standard strin# class!
+his is not a fundamental ty-e6 but it behaves in a similar way as fundamental ty-es do in
its most basic usa#e!
. first difference with fundamental data ty-es is that in order to declare and use ob/ects
@variablesA of this ty-e we need to include an additional header file in our source code2
Pstrin#Q and have access to the std names-ace @which we already had in all our -revious
-ro#rams than5s to the usin# names-ace statementA!
'2
&& my first strin#
Binclude PiostreamQ
Binclude Pstrin#Q
usin# names-ace stdN
int main @A
X
strin# mystrin# M G+his is a strin#GN
cout PP mystrin#N
return 0N
Y
O!erator#
1nce we 5now of the e0istence of variables and constants6 we can be#in to o-erate with
them! 7or that -ur-ose6 C?? inte#rates o-erators! 9nli5e other lan#ua#es whose
o-erators are mainly 5eywords6 o-erators in C?? are mostly made of si#ns that are not
-art of the al-habet but are available in all 5eyboards! +his ma5es C?? code shorter and
more international6 since it relies less on En#lish words6 but re4uires a little of learnin#
effort in the be#innin#!
Hou do not have to memori8e all the content of this -a#e! :ost details are only -rovided
to serve as a later reference in case you need it!
A##ignent 4A5
+he assi#nment o-erator assi#ns a value to a variable!
a M ;N
Arit%etic o!erator# 4 +B 7B CB 9B D 5
+he five arithmetical o-erations su--orted by the C?? lan#ua#e are2
'3
? addition
> subtraction
O multi-lication
& division
[ modulo
Re"ationa" and eE&a"it2 o!erator# 4 AAB FAB <B ;B <AB ;A 5
,n order to evaluate a com-arison between two e0-ressions we can use the relational and
e4uality o-erators! +he result of a relational o-eration is a <oolean value that can only be
true or false6 accordin# to its <oolean result!
We may want to com-are two e0-ressions6 for e0am-le6 to 5now if they are e4ual or if
one is #reater than the other is! Dere is a list of the relational and e4uality o-erators that
can be used in C??2
MM E4ual to
JM $ot e4ual to
Q reater than
P Eess than
QM reater than or e4ual to
PM Eess than or e4ual to
(ogica" o!erator# 4 FB GGB HH 5
+he 1-erator J is the C?? o-erator to -erform the <oolean o-eration $1+6 it has only
one o-erand6 located at its ri#ht6 and the only thin# that it does is to inverse the value of
''
it6 -roducin# false if its o-erand is true and true if its o-erand is false! <asically6 it returns
the o--osite <oolean value of evaluatin# its o-erand! 7or e0am-le2
J@; MM ;A && evaluates to false because the e0-ression at its ri#ht @; MM ;A is true!
J@C PM 'A && evaluates to true because @C PM 'A would be false!
Jtrue && evaluates to false
Jfalse && evaluates to true!
+he lo#ical o-erators && and ^^ are used when evaluatin# two e0-ressions to obtain a
sin#le relational result! +he o-erator && corres-onds with <oolean lo#ical o-eration
.$)! +his o-eration results true if both its two o-erands are true6 and false
otherwise! +he followin# -anel shows the result of o-erator && evaluatin# the
e0-ression a && b!
';
CONC(USION
Saraswati Com-uter Centre is the leadin# com-any of ,ndia in the business of Com-uter
+rainin#! +he com-any is located at :u5tsar! +he com-any is very fast #rowin# today!
While conductin# my trainin# , have learnt a lot re#ardin# C?? and many more detail of
which has already been #iven in this re-ort which may be read as -art of this re-ort! ,
have done my duty durin# trainin# with my heart & soul and can never for#et these
#olden days of my study! , assure that the thin#s which , learnt in the com-any will hel-
me in future career! Dence best of luc5 to Saraswati Com-uter Centre!
'C
'I'(IO*RAP,-
www!c??!com
www!saraswaticom-uteredu!com
www!#oo#le!com
www!wi5i-edia!com
www!alta>vista!com
www!yahoo>answers!com
www!en##123!com





'*