Sunteți pe pagina 1din 11

Masters of Engineering Thesis Proposal:

Explorations of Data Structures in a Programming System for Children



Thesis Supervisor: Professor Mitchel Resnick






















Tamara Stern

6.UAP Masters of Engineering Thesis Proposal
Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology
May 18, 2006

1

Abstract

Scratch is a media-rich programming language designed for children ages 10-16 by the Lifelong
Kindergarten group of the MIT Media Lab. One of the goals of Scratch is for kids to understand
basic programming concepts through their endeavors to create animated stories, video games,
music videos, and other multimedia and interactive pieces. Currently, the concept of data
structures is missing from Scratch. A handful of approaches have been tested in other
programming environments designed for kids, but are often underused because data structure
are presented in a very abstract way, with non- visible or manipulable representations. I
propose to design a more concrete and understandable representation of data structures to be
added to Scratch.

1. Introduction

Scratch is a media-rich programming language designed for children ages 10-16 by the Lifelong
Kindergarten group of the MIT Media Lab [1]. Its primary purpose is to enhance the
development of technological fluency at after-school centers in economically-disadvantaged
communities. Scratch is meant to be a programming environment in which kids can create
animated stories, video games, music videos, and other multimedia and interactive pieces. One
of the goals of Scratch is for kids to understand basic programming concepts through their
endeavors to create such projects.

Currently Scratch is missing a major core computer science concept: data structures. Without
data structures, kids are limited in the scope of projects they can create and the computer science
understanding that they will take away from Scratch. Although the area has been explored, there
has not been an understandable and useful model that fits within the Scratch paradigm.

My project will be focused on researching past attempts, exploring different methods, and
making suggestions for how to ultimately integrate simple data structures, such as lists and
strings, into Scratch. Particularly I want to research how to incorporate visible and manipulable
ways to represent these types of data. An example of a visible and tangible data structure
implementation that I will explore is a table of high-scores for a game. I also want to explore
how the user will interact with the data structures. Specifically, I will research the
representation, primitives, and functions for the data structures. I will prototype and test my
implementations at Computer Clubhouses (after-school centers for youth from low-income
communities) in order to see which idea works best with children. The ultimate goal is to
enhance Scratchs creative and educational worth through a richer functionality.

2. Background

The Lifelong Kindergarten group in the MIT Media Lab, led by Mitchel Resnick, specializes in
designing new technologies that expand the range of what people can design, create, and learn.
Specifically, the Lifelong Kindergarten group aims to introduce new technologies and activities
in classrooms, museums, and after-school centers, nurturing and studying the growth of
communities of playfully-inventive learners [2].

2

One of the efforts of the Lifelong Kindergarten
group is a collection of after-school centers called
Computer Clubhouses (see Figure 1). This effort,
funded by Intel and run in conjunction with the
Boston Museum of Science, is a worldwide
endeavor to bring technology to underserved
communities who would not otherwise have access
to technological tools and activities [2]. The culture
of computer clubhouse is one of self motivation, as
kids choose to work on projects in which they have a
personal interest.

Figure 1: Kids at a Computer Clubhouse.

2.1 Computer Clubhouse Cultures

Many informal cultures have developed in Computer Clubhouses. For example, a visual-design
culture has emerged where children are creatively expressing themselves with professional
graphic-design and image-processing software. Clubhouses have also developed a thriving
music-production culture.

These cultures are very productive in the clubhouse, but they do not engage kids in programming
as a form of self expression. The Lifelong Kindergarten group believes that it would be useful to
encourage kids to express themselves, create projects they care about, and understand core
computational ideas and concepts. In response to these beliefs, the group decided to build
Scratch, with the goal of introducing a programming culture at Computer Clubhouses [3].

2.2 Scratch Programming Environment

Scratch is a programming environment, built on the Squeak programming language, that enables
kids to create their own animations, videogames, interactive newsletters, music, simulations, and
interactive art. Scratch builds on the tradition
of other educational programming languages,
such as Logo [4], but takes advantage of new
computational ideas to make it easier to get
started with programming and extend the range
of what kids can create and learn [5].
Figure 2: Scratch layout
Blocks
Pallet
Scripting
Area
Object
Library
Stage
The basic layout of Scratch includes a stage
where programmed objects interact, an object
library, a scripting area for programs, and a
blocks pallet which contains a list of
programming blocks that can be used to build
scripts (see Figure 2).




3

3. Goals

The goal of this project is to design a data-structure model for Scratch that is conceptually
straightforward and easily used. The resulting model should exhibit these properties:
Visible. The data should be visible to the user as opposed to hidden in memory. I will
experiment with ways to have a concrete and clear representation of the data.
Manipulable. I want kids to feel comfortable directly manipulating the data, by editing
and touching it directly on the screen.
Tinkerable. Kids should feel able to tinker with data structures and the commands for
manipulating them by quickly and easily changing them and observing the outcome.
Useful. The addition of data structures into Scratch should provide children with new
opportunities for creative expression.
Easily Learned. The model should be easily learned such that a user can figure it out on
his or her own, as Scratch is meant to be used in informal settings outside of the
classroom.
Fits Within the Scratch Paradigm. The resulting model should make sense within the
current Scratch feel and fit with other Scratch metaphors.

4. Previous Work

Currently, there are several other programming languages, similar to Scratch, that are designed
specifically for kids. These programs have integrated data structures in various ways into their
software. Although some other programs might have some other goals, and their models might
not be fully applicable, there are lessons to be learned by looking at them. These examples will
serve as informative examples to which I will refer when beginning my own development of data
structures in Scratch.

4.1 Data Structures in Boxer

Boxer is a programming environment, designed in 1986, that is determined by two key principles
the spatial metaphor and nave realism [7]. Pieces of data are grouped together in zoom-able
boxes. Figure 3 shows how a phone
book database would be implemented
using Boxer. An overarching phone-
book object contains a list which stores
the entries of the phone book, a
function-1-key function which
changes the phone number of a person
in the phone book, and two variables,
number and name, which are provided
to the user in order to add people to the
phonebook.
Figure 3: Data Structures in Boxer
While this approach is very visual and
tangible, the Boxer metaphor is quite different from the Scratch metaphor. However, this
approach of zoom-able boxes in order to make all parts of a program completely tangible is
4

interesting, and will be considered during when deciding the proper way to integrate data
management into Scratch.

4.2 Data Structures in MicroWorlds

MicroWorlds, released in 1993, is an educational software tool meant to teach programming
concepts [6]. It is based on the Logo programming language which was pioneered by Dr.
Seymour Papert [12]. In MicroWorlds, there are two representations of a collection.

Abstract Forms. Lists are represented in an abstract way,
which can be instantiated and programmatically
manipulated by calling procedures, including methods
show, first, last, empty?, equal?, butfirst
(returns all but the first element), butlast (returns all but
the last element), and member?, in the command prompt
area below the stage.
Visual Forms. Strings are represented as words in a
textbox which is viewable on the stage area, as illustrated
in Figure 4. These textboxes may be manipulated by
directly typing into them.

MicroWorlds uses a parse command to convert the contents of a
textbox into an abstract list representation. For example, in order to extract a word from a
textbox, MicroWorlds requires the following command: pick parse <name of object>.
While this functionality is useful, the user is required to realize not only that the parse
command exists, but also that it is used to change words in a string to elements in a list, which is
necessary to extract a word from a sentence.
Figure 4: Text boxes in
MicroWorlds

Like the parse command, many of the interesting methods provided by MicroWorlds are
underutilized because of their lack of visibility. Consequently, many users are unable to explore
the possible functionalities of MicroWorlds without guidance.

While exploring ways to integrate data structures into Scratch, it will be essential to maintain the
visible and tangible nature of Scratch. Specifically, integrated list structures would necessarily
be presented on the screen and able to be directly manipulated, which differs from MicroWorlds
abstract list representation. Additionally, I will attempt to minimize the amount of overhead
necessary to perform simple tasks, such as picking a word from a sentence.

Other software has been created to help kids learn programming, such as StarLogo [8],
LogoBlocks [9], and Etoys [10]. Through my research in the next few months, I will continue to
explore previous attempts at integrating data structures into educational software applications.

5

5. Introducing Data Structures to Scratch

5.1 Data Structures in Scratch: Current Model

Currently, Scratch has one basic way to store data within the environment; specifically there are
variables that can store numbers. This functionality allows kids to maintain different types of
information in their Scratch projects. For example, a common use of Scratch is building
interactive games. The variable feature allows children to maintain the score of the current user.
Variables in Scratch may be created for all Sprites to use (global) or only for a particular sprite to
use (instance specific). In order to create a variable, the user clicks on the Variables Tab of the
Blocks Pallet. Here, they will see the option to create a new variable (for this sprite)
or a new variable (for all). Next, the user is asked to assign the variable a name, as
shown in Figure 5.

Figure 5: Creating a variablein Scratch



Once a variable is created, new blocks appear in the Blocks pallet that allow kids to
programmatically change the variable by a certain amount, set the variable to an amount, or
reference the variable somewhere in the Scripting Area. Figure 6 shows these blocks.

6

Figure 6: Variable blocks in the Blocks Pallet

Additionally, kids may drag and drop the variable reference (in this case, the score block) to the
Stage in order to view and manually manipulate the contents, as shown in Figure 7. This feature
allows the kids to actually see where information is maintained and have the control to change it
however they choose.

Figure 7: Ways to manipulate a variable manually in the Scratch Stage area. The
figure on the left displays the variable in an editable textbox, and the figure on the
right shows a slider that the user can control to change the value of the variable.


5.2 Data Structures in Scratch: New Directions

Maintaining and manipulating variables in Scratch has allowed kids to explore different uses of
Scratch, while also introducing them to the programming concept of maintaining state for a
Sprite. However, variables are limited to only storing single numbers. Other types of data
structures, such as collections or strings, would offer a new range of possible applications for
kids using Scratch. For example, one could imagine Scratch projects that include a high score
list, a batting average, or a phone number database (like the one presented in the Boxer example
above). It is the groups hope that integrating data structures into Scratch will provide a richer
environment to foster kids creativity.

7

For purposes of exploration, I will present two sample scenarios in which more involved data
structures would be useful.

5.2.1 Creating Interactive Mad Libs Example 1

In this example, I explore the blocks that would be necessary in creating a Mad Libs game in
Scratch. In particular, how a Scratch user could create random sentences from a collection of
nouns and verbs.

In Figure 8, Nouns, Verbs, and Sentences are collections which would be displayed on the
Stage Area. These collections would be created in the same way that a variable is created. Once
created, these collections could be manipulated in a similar way to the way in which
Microworlds collections are manipulated: either directly by the user (by typing into the text box)
or programmatically by using the Scratch blocks in the scripting area. Each collection would
have a set of corresponding blocks, including clear <collection>, add <string> to
<collection>, and random item from <collection> as shown in the example.

The code shown below takes random words from Nouns and Verbs and places them in
Sentence to create a random sentence.

Figure 8: Example of a Mad Libs implementation using data structures in
Scratch.


5.2.2 Maintaining a High Score List Example 2

Scratch users often create games, for which a players score is maintained. In this example, I
explore how a project would be created to maintain a list of high scores that are re-sorted each
time a new score is added to the list.
8


In the example shown in Figure 9, both High Score and Player are collections, while score
is a variable. All three of these objects will appear on the stage area of Scratch. After a game is
completed, the score and Player are added to High Scores on a separate line. Then, the
entire list of high scores is sorted by line, rather than by item. In other words, each line is treated
as a unified entity, and only the first word or number is used to sort the entities. This type of
line-by-line grouping lends itself to database-like interaction, where a line in the text box
represents a record in a database.

The new blocks presented in this example are add endline to <collection>, add
<item> to <collection>, <collection> (which can be treated as a string), and sort
<collection> <numerically / alphabetically / randomly> by <line /
word>.

Figure 9: Example of a high score list implementation using data
structures in Scratch.


6. Research Approach

6.1 Immediate Goals

Currently I am collecting and reading the existing literature on Scratch and other educational
software programs. In particular, the sources Ill begin with are: Mindstorms [11] by Seymour
9

Papert, Computers, Networks and Education [12] by Alan Kay and Some Reflections on
Designing Construction Kits for Kids [13] by Mitchel Resnick and Brian Silverman.

Additionally, I will continue to research ways in which other educational programming
languages integrate data structures into their systems. In particular, the programs Ill begin with
are: StarLogo [8], LogoBlocks [9], and Etoys [10].

6.2 Risks

When adding features to Scratch, there is often the trade-off of a more powerful software versus
a more understandable software. Often, the group has decided to make the software more
understandable and therefore omitting complicated, although potentially useful, features. There
is the possibility that adding data structures to Scratch might be considered too confusing for
Scratch users. This discovery would still be important as the group might reconsider decisions
made about Scratch. For example, objects in Scratch are currently made to be mostly
independent from one another. This decision was made to reduce confusion that might arise with
shared methods and variables. If this decision makes it difficult to integrate data structures in a
useful and clear way, the Scratch team might reconsider object independence.

6.3 Feasibility

I spent the summer after my freshman year, in MITs Undergraduate Research Opportunities
Program (UROP), working on the initial stages of development of Scratch. During that time, I
was able to participate in many discussions on the functionality of Scratch. I was also able to
add a variety of features to Scratch during my few months on the project. I am also familiar with
the Clubhouse environment as I have visited a few times. Additionally, I have taken a variety of
education courses at MIT, as well as spent time teaching children fulltime for a few months. I
feel that with all of this experience, I will be able to do the necessary research around adding
data structures in Scratch, implement prototypes, evaluate prototypes through user testing, and
make recommendations for ultimate integration of data structures into Scratch during the next
year and a half.

6.4 Timeline

2006 Spring - Read existing material on educational programming languages and
creative use of data structures.
- Explore other software applications and their integration of data structures.
- Explore ways to integrate data structures into Scratch.
- Complete 6.UAP writing requirement.

2006 Summer - Further explore ways to integrate data structures into Scratch.
- Determine feasibility, effectiveness, and usability.
- Re-familiarize self with programming language, Squeak.

2006 Fall - Prototype ideas for the chosen path.
- Test the prototypes with children at Computer Clubhouses.

10

2007 Spring - Analyze results and decide which implementations are best suited for the
Scratch programming environment and for the childrens needs.
- Complete thesis writing requirement.

7. Conclusion

In summary, adding data structures that are conceptually understandable and useful to
programming languages has proven difficult in the past. I propose to design a way to integrate
data structures into Scratch in a way that is comprehensible and effective. Through researching
past methods, prototyping, and user testing, I hope to design a system that is useful, clear, and
educational with the ultimate goal to enhance Scratchs creative and educational worth through a
richer functionality.

8. Acknowledgements

I would like to thank Mitchel Resnick and J ohn Maloney for their continued support and
guidance through my years with the Lifelong Kindergarten group at the Media Lab.

9. References

[1] Maloney, J ., Burd, L., Kafai, Y., Rusk, N., Silverman, B., and Resnick, M. (2004). Scratch:
A Sneak Preview, Second International Conference on Creating, Connecting, and
Collaborating through Computing. Kyoto, J apan, pp. 104-109.
[2] November (2005). http://llk.media.mit.edu/.
[3] Resnick, M., Kafai, Y., Maeda, J ., et al. (2003). A Networked, Media-Rich Programming
Environment to Enhance Technological Fluency at After-School Centers in Economically
Disadvantaged Communities. Proposal to the National Science Foundation (funded 2003-
2007). http://www.media.mit.edu/~mres/papers/scratch.pdf.
[4] November (2005). http://www.microworlds.com/.
[5] November (2005). http://weblogs.media.mit.edu/llk/scratch/.
[6] November (2005). http://el.media.mit.edu/logo-foundation/logo/.
[7] diSessa, A., Abelson, H. (1986). Boxer: A Reconstructible Computational Medium.
Communications of the ACM, vol. 29, no. 9, pp.859-868.
[8] November (2005). http://education.mit.edu/starlogo/.
[9] Begal, A. (1996). LogoBlocks: A Graphical Programming Language for Interacting with
the World. Unpublished Advanced Undergraduate Project, MIT Media Lab.
[10] Steinmetz, J . (2001). Computers and Squeak as Environments for Learning. In Rose, K.
and Guzdial, M. (eds.), Squeak: Open Personal Computing and Multimedia. Prentice Hall:
New York.
[11] Resnick, M., Silverman, B. (2005). Some Reflections on Designing Construction Kits for
Kids. http://web.media.mit.edu/~mres/papers/IDC-2005.pdf.
[12] Papert, S. (1980). Mindstorms: Children, Computers, and Powerful Ideas. Basic Books.
New York.
[13] Kay, A. (1991). Computers, Networks, and Education. Scientific American, vol. 265, no. 3,
pp. 100-107.
11

S-ar putea să vă placă și