Roll No.

(Write Roll Number from left side

exactly as in the Admit Card) Signature of Invigilator
Question Booklet Series X
PAPER–II Question Booklet No.
Subject Code : 22 (Identical with OMR
Answer Sheet Number)

COMPUTER SCIENCE AND APPLICATIONS

Time : 2 Hours Maximum Marks: 200

Instructions for the Candidates

1. Write your Roll Number in the space provided on the top of this page as well as on the OMR Sheet provided.
2. At the commencement of the examination, the question booklet will be given to you. In the first 5 minutes, you
are requested to open the booklet and verify it:
(i) To have access to the Question Booklet, tear off the paper seal on the edge of this cover page.
(ii) Faulty booklet, if detected, should be got replaced immediately by a correct booklet from the invigilator
within the period of 5 minutes. Afterwards, neither the Question Booklet will be replaced nor any extra time
will be given.
(iii) Verify whether the Question Booklet No. is identical with OMR Answer Sheet No.; if not, the full set is
to be replaced.
(iv) After this verification is over, the Question Booklet Series and Question Booklet Number should be entered
on the OMR Sheet.
3. This paper consists of One hundred (100) multiple-choice type questions. All the questions are compulsory. Each
question carries two marks.
4. Each Question has four alternative responses marked: A B C D . You have to darken the circle as
indicated below on the correct response against each question.
Example: A B C D , where C is the correct response.
5. Your responses to the questions are to be indicated correctly in the OMR Sheet. If you mark your response at
any place other than in the circle in the OMR Sheet, it will not be evaluated.
6. Rough work is to be done at the end of this booklet.
7. If you write your Name, Roll Number, Phone Number or put any mark on any part of the OMR Sheet, except
in the space allotted for the relevant entries, which may disclose your identity, or use abusive language or employ
any other unfair means, such as change of response by scratching or using white fluid, you will render yourself
liable to disqualification.
8. Do not tamper or fold the OMR Sheet in any way. If you do so, your OMR Sheet will not be evaluated.
9. You have to return the Original OMR Sheet to the invigilator at the end of the examination compulsorily and must
not carry it with you outside the Examination Hall. You are, however, allowed to carry question booklet and
duplicate copy of OMR Sheet after completion of examination.
10. Use only Black Ball point pen.
11. Use of any calculator, mobile phone, electronic devices/gadgets etc. is strictly prohibited.
12. There is no negative marks for incorrect answer.

15798 [ Please Turn Over ]

 X-3 22-II

COMPUTER SCIENCE AND APPLICATIONS

PAPER II
1. Which of the following is not correct? 4. Consider the following table in a relational
(A) Type 3 grammer includes type 0, type 1 and database:
type 2 grammers.
Last Name Designation Subject Course
(B) Type 0 grammer includes type 1, type 2 and
type 3 grammers. Roy Professor DBMS B.Tech.
(C) Type 1 grammer includes type 2 and type 3 Sen Associate OS M.Tech.
grammers. Professor
Roy Associate OS MCA
(D) Type 2 grammer includes type 3 grammer.
Professor
Pal Assistant Data B.Tech.
Professor structure
Assume the data in the table are valid, then the
candidate key of the table is
(A) {Last Name}
(B) {Last Name, Designation}
2. In what way ‘Cache Memory’ is different from
(C) {Subject, Course}
‘Main Memory’ for which ‘Cache Memory’ works
faster? (D) {Last Name, Course}

(A) Cache Memory doesn’t have addressing

system but Main Memory has.
(B) Cache Memory system is made of different
hardware material which makes it faster.
5. Divide and conquer policy is used in
(C) Cache Memory has different type of (A) Merge Sort
addressing system.
(B) Binary Search
(D) None of the above (C) Quick Sort
(D) All of the above

6. A thread is usually defined as a ‘Light weight

process’ because an Operating System (OS) maintains
3. Which of the following is not true? smaller data structures for a thread than for a process. In
relation to this, which of the followings is true?
(A) R  S  S  R (A) On per-thread basis, the OS maintains only
CPU register state.
(B) R  S  S  R (B) The OS does not maintain a separate stack for
each thread.
(C) R  S  S  R (C) On per-thread basis, the OS does not maintain
virtual memory state.
(D) On per-thread basis, the OS maintains only
(D) All of the above
scheduling and accounting information.
22-II X-4
7. Let the page fault service time be 10 mSec. in a 10. Which of the following represents the result of a
computer with average memory access time being query in a query tree?
20n Sec. If one page fault is generated for every 106 (A) root node
memory accesses, what is the effective access time for
(B) leaf node
the memory?
(C) intermediate node
(A) 21nSec.
(D) None of the above
(B) 30nSec.
(C) 23nSec.
(D) 35nSec.

11. Which of the following is not a monotonically

increasing grammer?
8. The number of edges in a regular graph of degree
d and n vertices is (A) Context-sensitive grammer

(A) n + d (B) Unrestricted grammer

(B) maximum of n, d (C) Regular grammer

(D) Context-free grammer
(C) nd
(D) (nd)/2

9. The cylomatic complexity of the following 12. Cosider the following table of arrival time and
program is: burst time for three processes P0, P1 and P2.
f = 1;
Process Arrival time Burst time
sum = 0;
for(i=1;i<=n;i++) P0 0 mSec. 9 mSec.
{
if(grade[i] <7) P1 1 mSec. 4 mSec.
f = o;
P2 2 mSec. 9 mSec.
sum = sum + grade[i];
} Pre-emptive shortest job first scheduling algorithm is
average = sum/n; used. Scheduling is carried out only at arrival or
if(averge > = 8 && f = = 1) completion of processes. What is the average waiting
printf (“Excellent”);
time for the three processes?
(A) 4
(A) 5·0 mSec.
(B) 6
(B) 4·33 mSec.
(C) 8 (C) 6·33 mSec.
(D) None of the above (D) 7·33 mSec.
 X-5 22-II
13. Which of the following transformation is 16. Suppose n processes, P1, ...., Pn share m identical
referred to as commutativity of projection and join? resource units, which can be reserved and released one
at a time. The maximum resource requirement of
(A) R  S  S  R
process Pi is Si, where Si > 0. Which one of the
following is a sufficient condition for ensuring that
(B) R  S  S  R deadlock does not occur?

(C)  L1 L 2  R S    L1 ( R) )  L 2 ( S )) (A) i, Si  m

C

(B) i, Si  n
(D) Both (A) and (B)
m
(C)  Si   m  n 
i 1
n
(D)  Si   m  n 
i 1

14. A Turning machine is an automation for

(A) Unrestricted grammers
(B) Context-free grammers
(C) Context-sensitive grammers
(D) Regular grammers 17. The postfix expression for the infix expression:
A + B * (C+D) / F + D * E is
(A) A + * BCD / F * + DE + +
(B) A * B + C D / F * DE + +
(C) ABCD + * F / + DE * +
15. Given the following function in C: (D) AB + CD + * F / D + E *
double f(double x, double Y)
{
int;
double a,b;
for (a=1, b=1, i=1; i<y; i++)
{
a* = x/i;
b+ = a;
}
return b; 18. OLAP is
}
For large value of Y, the best approximation of the (A) A dynamic synthesis of multi dimentional
return value of f is data

(A) xY (B) An analysis of multi dimentional data

(B) ex (C) A consolidation of large volumes of multi-
(C) ln(1+x) dimentional data

(D) yx (D) All of the above

22-II X-6
19. Which of the following testing is more 23. ALU of a computer can only perform ‘addition’.
comprehensive to test the components like hardware, Then how subtraction can be done in it and what type of
software and external interfaces? adder circuit is used in ALU (Arithmetic Logic Unit)?
(A) Unit testing (A) Additions using complements and serial
(B) Integration testing adder.
(B) Using different Algorithms and Carry Look
(C) System testing
ahead adder.
(D) Acceptance testing
(C) Addition using complements and Carry Look
ahead adder.
(D) Using ‘Carry and Save addition technique’
and ‘Carry Look ahead adder’.
20. In the IEEE 754 Floating point number format, 8
bit exponent field is represented by

 
(A) Excess –  2
n1
 1 Code

24. In multi-level index
(B) Excess –  2  1  Code
n
  (A) the datafile is ordered by the attribute that is
also the search key in the index file.
(C) Excess –  2  Code
n
(B) the datafile is ordered by an attribute that is

 
different from the search key in the index file.
(D) Excess –  2  1 Code
n1
 (C) an index structure consisting of two or more
where n = 8 . tires of records where an upper tier’s records
point to the associated index records of the
tier below.

21. Parallel algorithms are executed on (D) None of the above

(A) SISD system
(B) SIMD system
(C) MIMD system
25. Exact string matching algorithms are
(D) Both (B) and (C)
(A) KMP-algorithm
(B) Rabin-Karp algorithm
(C) Boyer-Moore algorithm
22. Consider the following recursive function:
(D) All of the above
int f(int x)
{
if(x==4)
return 2;
else return 2*f(x+1); 26. Primary functions of ALU (Arithmatic Logic
} Unit) are
The return value of f(2) is
(A) Arithmatic, Logical
(A) 2
(B) 4 (B) Arithmatic, Logical and Shifting
(C) 8 (C) Arithmatic, Shifting and Decoding
(D) 16 (D) Arithmatic only
 X-7 22-II

27. If f is _________ and g is _________ and non- 31. A file system with 300 GBytes uses a file
descriptor with 8 direct block address, 1 indirect block
increasing then h(x) = g (f (x) ) is convex. address and 1 doubly indirect block address. The size of
(A) Concave, Concave each disk block is 128 Bytes and the size of each disk
(B) Convex, Concave block address is 8 Bytes. The maximum possible file
size in this file system is
(C) Concave, Convex
(A) 3 Kbytes
(D) Convex, Convex
(B) 35 Kbytes
(C) 280 Bytes
(D) Dependent on the size of the disk

28. Which type of search can go on and on deeper

and deeper in the state space and we may get lost?
(A) A* 32. For the basic feasible solution to remain optimal
(B) Breadth First Search (A) all Cj–Zj values must remain zero.

(C) Depth First Search (B) no objective function co-efficients are allowed
to change.
(D) None of the above
(C) the value of objective function must not
change.
(D) All of the above

29. Which of the following technique is less costly in

designing ‘Control Unit’ of a computer?
(A) Hardware based Control Unit design
33. Physical Database design is the process of
(B) Micro-programming based Control Unit
(A) deriving the physical structure of the
design
database.
(C) Both Hardware and Software based designed
(B) creating a conceptual and external schemes
(D) None of the above
for the high level data model.
(C) analyzing overall data requirements.
(D) None of the above

30. A hash function h(k) = k mod 3 is given. What is

the number of collisions to store the following key
sequence? 34. Which of the following HTML tag is used to
15, 11, 34, 98, 10, 37, 51, 14 ,16, 47 create a numbered Lists?
(A) 7 (A) dL
(B) 2 (B) List

(C) 9 (C) uL

(D) 3 (D) oL
22-II X-8
35. For a K-nn classifier, as the value of K increases 39. The number of leaf nodes in a complete binary
tree of depth d is
(A) bias increases and variance decreases.
(A) 2d
(B) bias decreases and variance increases.
(B) 2d+1 + 1
(C) bias and variance both increases. (C) 2d–1 + 1
(D) bias and variance both decreases. (D) 2d + 1

40. Cryptographic hash function takes an arbitary

36. How ‘Cache-coherence’ problem can be block of data and returns
overcome in a multiprocesser system?
(A) a variable size bit string always.
(A) Using ‘MESI CACHE COHERENCY
PROTOCOL’ (B) a fixed size bit string always.
(B) Using LRU based system (C) any one of a variable or fixed size bit stream.
(C) Using Bus arbitration technique
(D) None of the above
(D) None of the above

41. Which one of the following is most appropriate

in case of functional testing?
37. A prolog program consists of a set of
(A) Test cases are designed on the basis of
(A) Facts internal structure of the source code.
(B) Test cases are designed on the basis of
(B) Rules
functionality ignoring internal structure of
(C) Clauses the source code.
(C) Test cases are designed on the basis of
(D) Conditions functionality and internal structure of the
source code.
(D) Test cases are designed on the basis of pre-
conditions and post-conditions.

38. In the back propagation algorithm, the gradient

of the error with respect to the weight vector is itself a
vector. What does the direction of the vector indicate? 42. Cosider a disk pack with 16 surfaces, 128 tracks
(A) It points in the direction of steepest decrease per surface and 256 sectors per track. 512 bytes of data
in error. are stored in a bit serial manner in a sector. The capacity
(B) It points in the director of steepest increase in of the disk pack and the number of bits required to
specify a particular sector in the disk are respectively.
the error.
(A) 256 Mbyte, 19 bits
(C) It indicates that component of the weight (B) 256 Mbyte, 28 bits
vector that results in the maximum error. (C) 512 Mbyte, 20bits
(D) None of the above (D) 64 Gbyte, 28 bits
 X-9 22-II
43. Which of the following is not a form of 47. Which one of the following is the formal process
functional testing? of deciding which attribute should be grouped together
(A) Boundary value analysis in a relation?
(B) Equivalence class testing
(A) Optimization
(C) Data flow testing
(D) Decision table based testing (B) Normalization
(C) Tuning
(D) None of the above

44. Which one is the use of Logic in AI?

(A) As a tool of analysis
48. The regular expression corresponding to the
(B) As a basis of knowledge representation
(C) As a language following finite automation is
(D) All of the above
a,b
a

(A) (a + b)*a
(B) (a + b) a*
(C) a* b* + a
45. The time complexity of a Turning machine to
recognize L over {0, 1} such that L = {0n 1n  n > 0} is (D) None of the above
of the order of
(A) n

(B) n2

(C) n3 49. Which database design method transforms

(D) n 4 DBMS independent conceptual model into DBMS
dependent model?
(A) Conceptual
(B) Logical
(C) Physical
(D) None of the above

46. If (rear = = maxsize –1)

rear = 0 ;
else
rear = rear +1;
is required in 50. Which of the following is not a UML diagram
(A) stack used in software development model?
(A) Class Diagram
(B) linear queue
(B) State Diagram
(C) circular queue (C) Data flow Diagram
(D) dequeue (D) Activity Diagram
22-II X-10
51. A CPU generates 32 bit virtual addresses. The 55. Which of the following ANN uses supervised
page size is 4KB. The processor has a translation look- learning?
aside buffer (TLB) which can hold a total of 128 page (A) Hopfield network
table entries and is 4-way set associative. The minimum (B) Self-organizing Map (SOFM)
size of the TLB tag is
(C) Simple Recurrent Network (SRN)
(A) 11 bits
(D) All of the above
(B) 13 bits
(C) 15 bits
(D) 20 bits

56. The register used as a working area in CPU is

(A) Program Counter
(B) Instruction Register
52. Properties of Bézier curves are: (C) Instruction Decoder
(A) It always passes through the first and last (D) Accumulator
control points.
(B) The boundary conditions at the two ends of
the curve are
P (O) = po
P (1) = pn
57. The production system SaA, A aB/a, B b
(C) It is a polynomial of degree one less than the represents
number of control points used.
(A) CFG
(D) All of the above
(B) Regular grammer
Which one is correct? (C) both CFG and regular grammer
(D) neither CFG nor regular grammer

53. Can boosting technique be applied on regression

problems? Can bagging be applied on regression 58. Given some data set, you are interested in
problems? identifying outliers. If you use clustering for this task
which approach would you prefer?
(A) No, No
(A) Hierarchical
(B) No, Yes (B) Over Lapping
(C) Yes, No (C) Partitional
(D) Partial
(D) Yes, Yes

59. Consider a relation R (A, B, C, D, E, F) having FDs

54. Consider a relation S (W, X, Y, Z) with FDs WX AB, BC  D, E  C, D  A. The candidate keys of
and YZ. The decomposition S1 (W, X) and S2 (Y, Z) is R are
(A) Lossless join. (A) AE, BE, DE
(B) Lossless join but not dependency preserving. (B) AEF, BEF, DEF
(C) Lossless join and dependency preserving. (C) AEF, BCF, BEF
(D) Dependency preserving but not loss join. (D) AEF, BDF, BEF
 X-11 22-II
60. What is a pattern vector? 64. Which of the following phase of a typical compiler
(A) A vector of measured features X = [X1,X2, ..., is not related to front end of that compiler?
Xn]T of an input example (A) Syntax analysis
(B) A vector of weights W = [W1,W2, ..., Wn]T is a (B) Lexical analysis
neural network (C) Code optimization
(C) A vector of outputs Y = [y1, y2, ..., yn]T of a (D) Semantic analysis
classifier
(D) All of the above

65. Consider a relation R (A, B, C) with the following

instance:

A B C
61. To implement top-down parsing the grammer
should be of the type 5 5 5
(A) LL (1) 5 7 8
(B) LL (2) 7 7 9
(C) LL (3)
Assume the above data are valid, which of the following
(D) All of the above functional dependencies must be false?
I. A  B
II. A  C
III. AB  C
62. Which of the ANN can be used to recover a clean (A) I only
version of a stored image given a noisy version of that (B) II only
image? (C) I and II
(A) Linear feedforward network (D) II and III
(B) Multi-layer feedforward network
(C) Simple recurrent network
(D) Hopfield network
66. Which of the following ANN would you use for
weather forecasting?
(A) Simple Recurrent Network (SRN)
(B) Self-Organizing Map (SOM)
63. Group 1 contains some CPU scheduling algorithms (C) Hopfield Network
and Group 2 contains some applications. Match entries in (D) Multi-layer feed forward network
Group I to entries in Group II.
Group I Group II
(P) Gang Scheduling (1) Guaranteed Scheduling
(Q) Rate Monotonic (2) Real-time Scheduling 67. Which one of the following is not correct?
Scheduling (A) If L1 and L2 are CFLs, then the language
(R) Fair Share (3) Thread Scheduling L1 V L2 is also a CFL.
Scheduling (B) If L1 and L2 are CFLs, then the language L1 L2
(A) (P) – (3) (Q) – (2) (R) – (1) is also a CFL.
(B) (P) – (1) (Q) – (2) (R) – (3) (C) If L1 is a CFL, then the language L = L1 * is
(C) (P) – (2) (Q) – (3) (R) – (1) also a CFL.
(D) (P) – (1) (Q) – (3) (R) – (2) (D) None of the above
22-II X-12
68. Consider a relation R (A, B, C, D, E) is in Third 74. Which of the following phase of the software
Normal Form. Which of the following functional development life cycle, regression testing is normally
dependencies must be FALSE? used?
(A) A, B  C (A) Design
(B) A, C  E (B) Coding
(C) A, B  D (C) Testing
(D) C, D  E (D) Maintenance

75. Which command is used to display the Unix

version 7?
69. Network Layer ‘Firewall’ works as a (A) uname–r
(A) Frame Filter (B) uname–n
(B) Packet Filter (C) uname–t
(C) Both Frame and Packet Filter (D) Kernel
(D) None of the above
76. Skolem function is used to remove
(A) 
70. In a tree, where the height of the two child subtrees (B) 
of any node differ by atmost one is called (C) variable
(A) Threaded Binary tree (D) None of the above
(B) AVL tree
(C) Complete tree 77. IP Sec is designed to provide the security at the
(D) Binary Search tree (A) Transport Layer
(B) Network Layer
(C) Application Layer
71. In A* algorithm heuristic values are admissible (D) Session Layer
where n is the current node, if
(A) h(n) > h*(n) 78. Let X, Y be two item sets and let Supp (X) denotes
(B) h(n) < = h*(h) the support of item set X. Then the confidence of the rule
(C) h(n) = constant Xy denoted by conf (xy) is
(D) None of the above
Supp ( X )
(A)
Supp (Y )
Supp (Y )
72. Parallel processing can occur in (B)
Supp ( X )
(A) the Instruction stream
(B) the Data stream Supp ( X  Y )
(C)
(C) Both (A) and (B) Supp ( X )
(D) None of the above
Supp ( X  Y )
(D)
Supp ( X )

73. Which of the following is a facility provided by 79. The expression ‘Deleyed Load’ is used in the
DBMS to assist the recovery process? context of
(A) Recovery Manager (A) Processor–Printer Communication
(B) Logging facilities (B) Memory–Monitor Communication
(C) Backup mechanism (C) Pipelining
(D) All of the above (D) None of the above
 X-13 22-II
80. What is the outpur of the following Prolog 85. In an indexed file organization, which is true?
program? (A) Index files must be direct access storage files.
fib (0,0). (B) Each key is associated with an index.
fib (1,1). (C) Each primary key value should not be changed
fib (N,R): N>1, during updation.
N1 is N–1, (D) All of the above
N2 is N–2,
fib (N1, R1),
fib (N2, R2), 86. Consider three CPU-intensive processes, which
R is R1+R2. require 10, 20 and 30 time units and arrive at times 0, 2
and 6 respectively. How many context switches are
?– fib (3,R)
needed if the operating system implements a shortest
(A) R = 0, 1, 1 remaining time first scheduling algorithm? Do not count
(B) R = 1, 1, 2 the context switches at time zero and at the end.
(C) R=2 (A) 1
(D) R=3 (B) 2
(C) 3
81. In object-oriented programming language, (D) 4
‘generalization’ is implemented using
(A) Overloading
87. Consider the following function:
(B) Encapsulation
int f()
(C) Data Hiding
{
(D) Inheritance int i, s;
s=0;
82. Consider a machine with 64 MB physical memory for (i=0;i<5; i++)
and a 32-bit virtual address space. If the page size is 4KB,
{ if (i%3==1)
what is the approximate size of the page table?
(A) 16 MB s+=i;
(B) 8 MB else
(C) 2 MB s++;
(D) 24 MB .}
return s;
83. In JavaScript which of the following is used to }
declare variable?
The return value of f() is
(A) function
(B) var (A) 5
(C) dim (B) 7
(D) None of the above (C) 6
(D) 8
84. Fourth Normal Form (4NF) is concerned with
dependencies between the elements of compound keys
composed of 88. A proxy firewall filters at
(A) one attribute (A) Physical Level
(B) two attributes (B) Data Link Layer
(C) three or more attributes (C) Transport Layer
(D) None of the above (D) Application Layer
22-II X-14
89. In UML the diagram used in describing the system 93. A software requirement specification is
behaviour during execution is
I. a contact between client and software developers
(A) State Chart Diagram
and clients specifying what the developers will
(B) Deployment Diagram
develop for the clients.
(C) Implementation Diagram
(D) All of the above II. a specification of the features that the software to
be developed must have.
III. a specification of the resources along with man
power that will be required for the development of
the software.
(A) I only
(B) II only
(C) I and II only
90. Basic components of IOT consist of
(A) Sensor Nodes (D) I, II and III
(B) Sensor Nodes and Gateways
(C) Gateways and Cloud
(D) Sensor Nodes, Gateways and Cloud

94. Y 1
Original position
2 3
( X
91. Which of the following is not the metric for object- / /
2 3
oriented testing?
Reflected position
(A) Number of key classes
/
(B) Fan In (FIN) 1
(C) Lack of Cohesion in methods
The above reflection is accomplished with the
(D) Public Access to Data members (PAD)
transformation matrix:

1 0 0 
(A) 0 1 0 
 
0 0 1 

 1 0 0 
(B)  0 1 0 
 
 0 0 1 
92. Which Protocol assigns IP address to the client
connected in the internet?  1 0 0
(A) DHCP (C)  0 1 0
 
(B) IP  0 0 1 
(C) RPC
(D) None of the above (D) None of the above
 X-15 22-II
95. In the Slow-start algorithm, the size of the 98. Traffic intensity is given by, where L = n of bits
Congestion window increases _______ until it reaches a in a packet, a = Average rate, R = transmission rate.
threshold value.
(A) Multiplicatively (A) LR
(B) Exponentially a
(C) Additively (B) L
(D) Randomly Ra
(C) La
R
(D) R
La

99. You want to enable both ARPA and SNAP

96. In a computer a negative fixed point binary number
encapsulation on one router interface. How do you do
is represented by the format
this?
(A) Sign bit – Magnitude
(A) The interface can handle multiple
(B) Sign bit – 1’s Complement encapsulation types with no extra
(C) Sign bit – 2’s Complement configuration.
(D) Only 2’s Complement (B) Assign two network numbers, one for each
encapsulation type.
(C) Enable Novell-ehter to run multiple
encapsulation types.
(D) Both APRA and SNAP are enable by default,
so you don’t have to configure anything.
97. For a 10 Mbps Ethernet Link, if the Length of the
packet is 32 bits, the transmission Delay is (in 100. Which type of Ethernet framing is used for
Microseconds) TCP/IP and DEC net?
(A) 3·2 (A) Ethernet 802·3
(B) 32 (B) Ethernet 802·2
(C) ·32 (C) Ethernet II
(D) 320 (D) Ethernet SNAP
22-II X-16
ROUGH WORK
 X-17 22-II
ROUGH WORK
22-II X-18
ROUGH WORK
 X-19 22-II
ROUGH WORK
22-II X-20
ROUGH WORK