Fun in CS2

∗

Amalia Duch, Jordi Petit, Enric Rodr´ıguez-Carbonell and Salvador Roura

Departament de Llenguatges i Sistemes Inform`atics,

Universitat Polit`ecnica de Catalunya, Barcelona, Catalonia, Spain

Keywords:

Data Structures, Algorithms, Computer Game, Computer Game’s Player, Programming.

Abstract:

We report our experience of including the implementation of a player of a computer game as a programming

project in a CS2 course focusing in data structures and algorithms. Every semester, an instructor designs the

rules of a game, prepares its visualization, and implements an elementary player with a very simple strategy.

The game is then delivered to students who, as a ﬁrst step in order to pass the project, must program a player

that wins the elementary player. Then, a tournament begins among all the students with accepted players. At

every round of this tournament a player is eliminated until just one player, the champion, survives. Grades for

this assignment are computed automatically and increasingly with respect to the round where students have

been eliminated. The result is a fun and very motivating programming experience for our students.

1 INTRODUCTION

As professors of algorithmic subjects, we face nowa-

days that, frequently, our students are poorly moti-

vated by our class matters but pretty hooked to com-

puter games. So, following the saying “If the moun-

tain will not come to Muhammad, then Muhammad

must go to the mountain”, we have decided to intro-

duce computer games in our data structures and algo-

rithms courses.

We have done so by means of a programming ac-

tivity consisting in the implementation of a strategy

for a player of a computer game. The goal is that the

programming of this player should involve the imple-

mentation of some of the algorithms and data struc-

tures taught during the course, and to motivate for the

self study of others.

Here, to program a player consists in designing

strategies to move some tokens, having complete in-

formation of the map in which the tokens are ly-

ing and about the rest of the players. In fact, each

player controls a set of conﬁgurations that will inter-

act with the game by reading information about the

game world and inform of its moves to interact with

it at each turn.

Note that games are meant to be played by pro-

grams, and to be seen by humans. There is no human

interaction during the games. The program to con-

∗

Work supported by the Generalitat de Catalunya’s

project ALBCOM (ref. 2009 SGR 1137).

trol the intelligence of a player is written before each

match starts.

The program submitted by a student (a player) will

be faced to the programs submitted by other students,

and will get better or worse score depending on the

way it is planned, i.e., how good is in practice its

programmed strategy. Unlike most exercises of the

course, this project is open: it is not about solving

a speciﬁc problem, but to prepare a strategy, to study

the strategies of theopponentplayers, and, if required,

adjust or reprogram the own strategy.

The idea was inspired by the ICPC-Challenges or-

ganized by ACM during the World Finals of their

annual ACM International Collegiate Programming

Contest

. A similar initiative is the AI Challenge

The performance of this lab assignment encour-

ages, among other skills, algorithmic programming

aspects (implementing their own strategy), program-

ming related to artiﬁcial intelligence (controlling the

heuristics of the agents that appear in the game) and

even good programming habits, because the players

are expected to change over the project duration to

adapt to their peers’ strategies, and thus, the students

with the most ﬂexible and maintainable code will be

beneﬁted.

There are several other initiatives that include au-

tomatic evaluation and assessment systems, digital

and web support, competitive and problem oriented

See http://icpc.baylor.edu.

See http://aichallenge.org.

437

Duch A., Petit J., Rodríguez-Carbonell E. and Roura S..

Fun in CS2.

DOI: 10.5220/0004389604370442

In Proceedings of the 5th International Conference on Computer Supported Education (CSEDU-2013), pages 437-442

ISBN: 978-989-8565-53-2

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

learning, among others (Gim´enez et al., 2012; Douce

et al., 2005; Saikkonen et al., 2001; Joy et al., 2005;

Cheang et al., 2003; Kosowski et al., 2007; Kur-

nia et al., 2001; Revilla et al., 2008). But, as we

will show, thanks to the enthusiasm that video games

rouse amongst students, the experience presented in

this paper is particularly engaging.

2 CONTEXT AND GOALS

This project is included as part of the “Data Structures

and Algorithms” course of the Facultat d’Inform`atica

de Barcelona and the “Algorithmics” course at the

Facultat de Matem`atiques i Estad´ıstica (both at the

Universitat Polit`ecnica de Catalunya). These courses

(with minor differences between them) start by intro-

ducing the concepts of algorithm analysis, together

with the required basic mathematical tools. Then,

these bases are used to study and analyze various im-

plementations of classic and essential algorithms and

data structures (sorting and searching, divide and con-

quer, dynamic programming, hashing, balanced bi-

nary trees, graph traversal, shortest paths, among oth-

ers).

These courses intend to combine basic theoreti-

cal aspects of algorithm design and analysis together

with several programming features. They consist of

6 ECTS

that involve two hours of theory classes, one

hour of classes of problems and one hour of labora-

tory classes per week.

Inside such courses, the educational objectives of

the game are the following:

• To provide an original and motivating environ-

ment to facilitate the integration of generic and

technical skills.

• To supply an active learning activity that improves

the learning of programming, data structures and

algorithms.

• To encourage the development of competencies

such as the design and analysis of effective, ef-

ﬁcient, collaborative and/or competitive strategies

to obtain well-deﬁned objectives.

• To introduce didactic materials in new formats us-

ing new technologies (web, social networks, fo-

rum, graphics, etc.).

ECTS stands for European Credit Transfer System,

which is a standard unit for comparing the study attainment

of students of higher education across the European Union.

One academic year corresponds to 60 ECTS-credits that are

equivalent to 1500–1800 hours of study, i.e., 25–30 hours

of study per credit.

• To promote the programming and the design of

efﬁcient algorithms, improving the integration of

practical and theoretical learning.

• To offer tools that allow students to strengthen

competencies such as the gradual and continu-

ous improvement and/or overcoming of previous

work at both personal and social levels.

• To promote new evaluation mechanisms for the

skills of programming and design of efﬁcient al-

gorithms.

3 DEVELOPMENT

The development of this activity has ﬁve main phases:

preparation, distribution, qualiﬁcation, tournament,

and grading.

Preparation. A newgame is created by one or more

instructors of the course; some actual games are pre-

sented later, in Section 4. A game consists of the fol-

lowing elements:

• The documentation gives a basic overview of the

game, describes its goals and rules, and presents a

tutorial on how to use it.

• The interface of the game is the C++ API that

users will use to write their own strategies. It basi-

cally includes functions to query the current state

of the board and functions to order changes to the

agents in the board.

• The implementation of the game includes all nec-

essary coding to make the game work, such as

reading the initial boards, handling the rules, up-

dating the score, etc. It also includes the viewer of

the game, that is, the tool that will be used to see

games in a graphical way. For portability reasons,

this viewer is programmed in Javascript.

• An elementary player called the dummy, which

implements a fairly simple game strategy.

Distribution. All the above elements (including the

source ﬁles) are given to the students, with the excep-

tion of the dummy player, for which only the object

ﬁles are distributed. With this material, all students

can design, implement, run, debug and view their own

games at home as often as needed. In addition, stu-

dents are given access to an online web server that

will handle their submissions and games.

Moreover, we try to motivate students as much

as possible by designing advertisement posters of the

game. We hang them on the walls of classrooms and

campus squares.

CSEDU2013-5thInternationalConferenceonComputerSupportedEducation

438

Qualiﬁcation. After distribution, students are given

three weeks to work on their own designing a strat-

egy and writing a program for their players. The only

requirement for their players to qualify is to consis-

tently beat the dummy player. This requirement is

necessary to ensure that a minimal coding effort is

done by each student; otherwise even an empty strat-

egy would qualify. In order to prove that they beat the

dummy player, students submit their programs to our

online judge, which checks that in four random games

against dummy players, the submitted player always

wins. Students have an unlimited number of oppor-

tunities to beat the dummy player within the three

weeks deadline.

Tournament and Grand Final. The most exciting

phase of the game consists in the play-off tournament,

which starts after the previous deadline with all the

qualiﬁed players. Over a couple of weeks, the game’s

website automatically takes care of the tournament,

handling matches among subgroups of students (typ-

ically four) and performing the necessary rounds to

give a ranking to all the players by successively elim-

inating the “worst player” of each round. All the

games and results disputed during the tournament are

publicly visible in the game website.

From round to round, students may still submit

new strategies that will replace their formers ones,

in order to try to react to the actions of their mates.

These replacement players must, of course, also beat

the dummy.

When only a few (16, say) players stand, a grand

ﬁnal is organized in the conference room of the Fac-

ultat d’Inform`atica de Barcelona to discover the ﬁnal

champion. The surviving players are grouped into

semi-ﬁnals, and the winners of each semi-ﬁnal dis-

pute the last round, where only one of them survives.

During the ceremony, the programmers of these best

players are invited to give a public short speech ex-

plaining how their players work and how they are pro-

grammed.

Grading. The total grade for our courses is a num-

ber between 0 (nothing) and 10 (perfect). Out of these

10 points, 9 come from standard evaluation systems

(tests, exams, lab assignments). Then, all the students

who beat the dummy before the deadline obtain one

point. In addition, and according to their ﬁnal ranking

in the tournament, each student receives up to an extra

point to be added to his ﬁnal grade (for a maximum

of 10, of course). This is computed proportionally to

the time their player is still in the game, thus the ﬁrst

eliminated player gets no extra point while the cham-

pion gets one extra point.

4 ACTUAL GAMES

Over the past few semesters, we have created several

games, trying our best to get attractive and enjoyable

activities. All the games feature a board where several

agents controlled by four different players interact for

several rounds to get a ﬁnal score. See Figure 1 for

some examples.

PacMan. This is an adaptation of the classical arcade

game for four players, where each player controls

his own pacman and three or four ghosts. Nor-

mal pills eaten by pacmen score points, and power

pills temporally boost pacmen to move faster and

to eat ghosts, in order to get extra points and

weaken adversaries.

Battle Royale. Each player controls several knights

and peasants. As they move to adjacent cells,

peasants colonize cells. Knights capture (kill) ad-

versary peasants to convert them to their team and

probabilistically ﬁght adversary knights. For each

player, the score is the ﬁnal number of colonized

cells. The name of the game comes from the fa-

mous Japanese cult ﬁlm, just for the large number

of kills during each game.

Apocalypse Now. In this adaptation of the Vietnam

war ﬁlm, the goal of each player is to con-

trol as many check points as possible. To do

so, each player can use soldiers, helicopters

and parachutists, who have different mobility at-

tributes depending on the terrain they move (jun-

gle, ﬁeld, water or mountain) and have different

kinds of attack (body to body or napalm).

Tron. This is a variation of the classic game and ﬁlm

where each player controls four light cycles that

should not crash with the preﬁxed walls in the

maze nor with trails of light that all light cycles

leave behind. In case of a crash, the light cycle

and its trail disappear from the board. The winner

is the last player that survives.

Dragon Ball. Inspired by the famous animated se-

ries, in this game each player must control Son

Goku in his quest to collect dragon balls and bring

them to hoi-poi capsules. In the process, play-

ers can launch kame hame attacks to other players

and use kinton clouds to move faster.

In all the games, the board is organized as a col-

lection of cells that induce a graph. At each round,

the movements of the agents are governed by each

player strategy and invoked through the game API.

Each player decides its movements for the next round

independently of the other players, and there is a ran-

domization process by which possible collisions are

resolved.

FuninCS2

439

Battle Royale

PacMan

Tron

Dragon Ball

Figure 1: View of some games.

5 THE ONLINE SYSTEM

Students use a website to submit their solutions, check

whether their programs beat or not the dummy, re-

quest training games with other players (so that they

do not need to exchange their source code to play

games among them) and track the results of the

tournament games, which presumably determines the

quality of their programs. This website is largely in-

dependent of the game, so is reused each semester.

To dispute the games, our system uses the facili-

ties of the Jutge.org infrastructure, which is an educa-

tive online programming judge developed at our Uni-

versity and open to all users (Gim´enez et al., 2012).

In brief, Jutge.org offers a highly scalable and secure

system that processes submissions to programming

assignments and offers a verdict on their behavior.

The techniques to dispute our types of games must

however extend the security measures used in simi-

lar online judges (Foriˇsek, 2006), because not only

they must prevent players to abuse or tamper the sys-

tem, but they also must prevent to cheat other players.

To do so, our system disputes the games by conﬁn-

ing each player to its own process with its own user.

Moreover, there is an additional process that medi-

ates in their communication (to check that they obey

the rules of the game) and supervises the player pro-

cesses (to check that none exceeds memory or time

limits and to handle this case if needed).

While these levels of security can seem paranoid

at ﬁrst, they are rather necessary, as some students are

able to devise contrived ways to try to cheat. For in-

stance, during the last semester, a group of students

has used knowledge on the way that the parameters

are pushed on the stack of function calls to locate

the memory positions that encode the walls of the

maze, and so could change them to their advantage.

While this hack worked with the distribution of code

we give to students and could amaze (or even terrify)

some of their opponents, it did not have any conse-

quences inside the server that plays the ofﬁcial tour-

nament thanks to not sharing the memory space.

In addition, our system interacts with the well-

known JPlag service to detect (and specially discour-

age) plagiarism (Prechelt et al., 2007). This could be

an important issue (it is not), because none of the sub-

mitted programs will normally be read.

As an example of the magnitude of the tourna-

ments, in the semester PacMan was played, we had

143 students that submitted 2929 programs, 749 of

them beating the dummy player. A total number of

7761 games were disputed and 2.2 GB of data storage

was needed.

CSEDU2013-5thInternationalConferenceonComputerSupportedEducation

440

I liked the game. It was easy to beat the

dummy player.

I would rather spend my

time studying theory than

programming my player.

To program a player is

much more motivating to

me than traditional pro-

gramming projects.

Competing against my

classmates in the game

motivates me.

I have taken the time to im-

prove my player after win-

ning the dummy player.

The documentation of the

game is right to me.

I would rather prefer this

course without game.

Figure 2: Results of the survey for the PacMan game (1: Strongly disagree → 5: Strongly agree). Values are in percentages.

6 SUPPORT

It may be clear now that, besides the investment of

building an online system to process submissions, the

development of this activity repeatedly calls for some

effort of the course instructors.

This support starts at the preparation phase (we es-

timate that this involvesabout 40 hours of work), con-

tinues monitoring the game and reacting in the case

of need (we estimate this involves about 10 hours of

work) and ﬁnishes preparing the grand ﬁnal.

Indeed, as most participants and many of their

classmates look forward to the grand ﬁnal of the

game, this is an important event that is well prepared

in advance: slides, game matches, videos, viewers,

etc. The ﬁnal is attended by many students and

faculty staff. It is worth commenting also that in

order to increase the students excitement and expec-

tation about the grand ﬁnal, we design a new visual

interface (as attractive as possible) together with a

stylish soundtrack. A video recording the grand ﬁnal

of the Battle Royale game can be found at http://

media.ﬁb.upc.edu/ﬁbtv/streamingmedia/view/2/209.

At the moment, this is the video with most views

among all videos posted in our computer science

school.

In addition, one has also to add up the time to give

support to the students that request help from their

instructors. Some of them require individual or group

ofﬁce hours.

Figure 3: Results for the question TimeI have spent working

on my player.

0 0.2 0.4 0.6 0.8 1

Figure 4: Scatter plot of ﬁnal grade (from 2 to 10) accord-

ing to rank in the PacMan tournament (from 0 [worst] to 1

[best]).

7 RESULTS

The game activity has been implemented as described

in the latest ﬁve semesters. It got a large percentage of

participation (always more than 90%) and almost all

FuninCS2

441

students who tried it could enter a player into the tour-

nament, i.e., their player could win the dummy and

therefore, they passed the project. Moreover, since

to keep alive during the tournament provides extra

points for the course’s grade, this activity really helps

several students to pass the course (about a 20% of

students pass the whole course because of this extra

point).

In order to support these statements, last semester

we applied a survey to our students. Figures 2 and 3

show the results of each of the issues on which we

consulted students. The range of answers was from

1 (strongly disagree) to 5 (strongly agree). From the

surveys, we can infer that almost all students enjoy

the game activity, that they do not ﬁnd too many prob-

lems in beating the dummy player, and that they pre-

fer it to more standard assignments. Also, they like

to compete against each other. Therefore, we have

corroborated that presenting the project as a game has

had a great impact on teaching. Moreover, this prac-

tice model is very attractive and motivatingto our stu-

dents.

On the other hand, one could think about some

correlation between the ranking of the students in the

tournament and their ﬁnal grades in the course. Fig-

ure 4 shows this information, and the Spearman’srank

correlation concludes that although the correlation is

weak (Rho = 0.38), the correlation is statistically sig-

niﬁcant (p-value = 2.757e− 05).

As said, during the grand ﬁnal, the programmers

of the best players explain their strategies. According

to their explanations, in a few cases it is possible to

reach the grand ﬁnal with just a very simple player,

but in general, winners implement smart and compli-

cated strategies, with sophisticated algorithms.

8 DISCUSSION

Along this work we have described our experience in-

troducing programming computer strategies for com-

puter games into typical CS2 courses. We have shown

evidence on how this kind of programming activity is

fun and highly motivating to computer science and

mathematics students. This motivation also encour-

ages professors, and thus facilitates a pleasant work-

ing environment.

As a weak point of such a project, one can think

that students seem to be so hooked that they spend

more than the recommended hours to program and

improve their players, and this in detriment of the

hours they should dedicate to other parts of the course

and to other courses. But, when surveyed about this

issue, they did claim that they dedicated a big amount

of time and work, but that this was mostly in their free

personal time, which otherwise they would not have

dedicated to study.

Overall, we consider it such a successful experi-

ence that we want to continue improving and spread-

ing. Among the ideas we have to go further, it is worth

mentioning that the games we have created so far are

very competitive. So, in the future we would also

like to develop collaborative games in which players

should help each other in some way. On the other

hand, we want to extend our online system to offer

these games to the general public. A systematization

of the website and of the design process of new games

is under current development.

ACKNOWLEDGEMENTS

We thank Omer Gim´enez and Mario G. Munz´on for

all their enthusiasm, ideas and programming time.

REFERENCES

Cheang, B., Kurnia, A., Lim, A., and Oon, W.-C. (2003).

On automated grading of programming assignments

in an academic institution. Computers & Education,

41(2):121–131.

Douce, C., Livingstone, D., and Orwell, J. (2005). Auto-

matic test-based assessment of programming: A re-

view. ACM Journal on Educational Resources in

Computing, 5(3).

Foriˇsek, M. (2006). Security of Programming Contest Sys-

tems. In Dagiene, V. and Mittermeir, R., editors, In-

formation Technologies at School, pages 553–563.

Gim´enez, O., Petit, J., and Roura, S. (2012). Jutge.org: An

educational programming judge. In Proc. of the 43rd

ACM Technical Symposium on Computer Science Ed-

ucation (SIGCSE-2012), pages 445–450. Association

for Computing Machinery.

Joy, M., Grifﬁths, N., and Boyatt, R. (2005). The BOSS on-

line submission and assessment system. ACM Journal

on Educational Resources in Computing, 5(3).

Kosowski, A., Malaﬁejski, M., and Noinski, T. (2007). Ap-

plication of an online judge & contester system in aca-

demic tuition. In ICWL’07, pages 343–354.

Kurnia, A., Lim, A., and Cheang, B. (2001). Online judge.

Computers & Education, pages 299–315.

Prechelt, L., Malpohl, G., and Philippsen, M. (2007). JPlag:

ﬁnding plagiarisms among a set of programs.

Revilla, M., Manzoor, S., and Liu, R. (2008). Competitive

learning in informatics: The UVa online judge experi-

ence. Olympiads in Informatics, 2:131–148.

Saikkonen, R., Malmi, L., and Korhonen, A. (2001).

Fully automatic assessment of programming exer-

cises. ACM SIGCSE Bulletin, 33(3):133–136.

CSEDU2013-5thInternationalConferenceonComputerSupportedEducation

442