CS61B Project 3: BYOW
Introduction
Skeleton Code
Phase 1: World Generation
Tileset and Tile Rendering Starting Your Program Requirements
Design Document
Design Document Guidelines
Phase 2: Interactivity
UI (User Interface) Appearance UI Behavior
Saving and Loading
Interacting With Input Strings
Ambition Score
1080 Point Primary Features 360 Point Secondary Features
Requirements Summary Submission and Scoring Details Checkoff Script and Form Autograder Details
Phase 1 Grader
Phase 2 Grader
Each assignment will have an FAQ linked at the top. You can also access it by adding “/faq” to the end of the URL. The FAQ for Project 3 is located here.
Introduction
In Project 3, you will create an engine for generating explorable worlds. This is a large design project that will require you and one partner to work through every stage of development from ideation to presentation. The goal of this project is to teach you how to handle a larger piece of code with little starter code in the hopes of emulating something like a product development cycle. In accordance with this, the grading of this project will be different from other projects. Since there is no notion of “the correct answer” when it comes to world design and implementation, you will be assessed much like a performance review you might receive at an internship or job in addition to a very general autograder. While this means you will be graded slightly subjectively, we promise to be pretty nice bosses and will respect you as any boss should respect their hardworking employees. Please talk to us if you feel the grading scheme feels unfair.
This project will require you a great deal of exploration and experimentation. Searching the web for answers (not solutions from past semesters) should be a regular activity
throughout this process. Please know that there are no right and wrong answers, as this is a very open-ended project. However, there are some implementations and ideas that are better than others. It is ok and expected that you will go through several iterations before settling on something that you deem good. That is, this project is about software engineering.
You’re not required to use any of the fancy data structures or concepts from class (A*, MSTs, Disjoint Sets, etc.). This project is about software engineering, not about data structures or algorithms. The data structures and algorithms we’ve learned about in class will make your code significantly simpler and more efficient, but please don’t use things just because we learned about them in class. Only use these tools if you feel comfortable using them in your implementation.
A video playlist (from Spring 2018) discussing tips for working on this project can be found at this link. Note that the API has changed slightly, but the basic ideas are all still true. Slides for these videos can be found at this link.
There are several key deadlines for this assignment:
Phase 0 – Team formation: You must submit the Project 3 Partnerships Form by Wednesday, April 5th at 11:59 PM PST. You will not be able to change your partner later. Read and understand the partnership guidelines before starting the assignment.
Phase 1 – World Generation: Due on Gradescope by 4/14 at 11:59 PM.
Phase 2 – Interactivity: Due on Gradescope by 4/24 at 11:59 PM.
You cannot submit Phase 2 of the project late, as it will be graded during a lab checkoff with a TA (including 47b students). While in theory you could submit Phase 1 and the supporting labs (Lab 12 and Lab 13) late, Phase 2 of the project builds upon these assignments, so it is unlikely that you will be able to submit these assignments late and still complete Phase 2 on time.
For convenience, here are the form links for this project:
Partnerships Form Phase 1 Reflection Phase 1 Checkoff Phase 2 Reflection Phase 2 Checkoff
Now on to the assignment spec!
Your task for the next few weeks is to design and implement a 2D tile-based world exploration engine. By “tile-based”, we mean the worlds you generate will consist of a 2D grid of tiles. By “world exploration engine” we mean that your software will build a world, which the user will be able to explore by walking around and interacting with objects in that world. Your world will have an overhead perspective. As an example of a much more sophisticated system than you will build, the NES game “Zelda II” is (sometimes) a tile based world exploration engine that happens to be a video game:
The system you build can either use graphical tiles (as shown above), or it can use text based tiles, like the game shown below:
We will provide a tile renderer, a small set of starter tiles, and the headers for a few required methods that must be implemented for your world engine and that will be used by the autograder. The project will have two major deadlines. By the first deadline, you should be able to generate random worlds that meet the criteria below. By the second deadline, a user should be able to explore and interact with the world.
The major goal of this project is to give you a chance to attempt to manage the complexity that comes with building a large system. Be warned: The system you build probably isn’t going to be that fun for users! Three weeks is simply not enough time, particularly for novice programmers. However, we do hope you will find it to be a fulfilling project, and the worlds you generate might even be beautiful.
Skeleton Code
As you did in Project 2B, use in your group repo to pull the skeleton code. The skeleton code contains two key packages that you’ll be using:
provides some basic methods for rendering, as well as basic
code structure for tiles, and contains:
– contains rendering-related
TERenderer.java
methods. the world.
– the type used for representing tiles in – a library of provided tiles.
IMPORTANT NOTE: Do NOT change TETile.java’s
field or method as it may lead
git pull skeleton main
byow.TileEngine
byow.TileEngine
TETile.java
Tileset.java
character()
程序代写 CS代考 加微信: cstutorcs
to bad autograder results.
RandomUtils.java
The other package contains everything unrelated to tiles. We recommend that you put all of your code for this project in the package, though this not required. The package comes with the following classes:
– Handy utility methods for doing randomness related things.
– How the user starts the entire system. Reads command line arguments and calls the
appropriate function in .
– Contains the two methods that allow interacting with your system.
provides two methods for interacting with your system. The first is
. This method takes as input a series of keyboard inputs, and
returns a 2D array representing the state of the universe after processing all the key presses provided in input (described below). The second is
. This method takes input from the keyboard, and draws the result of each
keypress to the screen. Lab 12 covers how to render tiles, and Lab 13 covers how to get user input.
This project makes heavy use of , which is a package that has basic graphics rendering capabilities. Additionally, it supports user interaction with keyboard and mouse clicks. You will likely need to consult the API specification for at some points in the project, which can be found here.
Engine.java
Engine.java
byow.Core.Engine
public TETile[][] interactWithInputString(String input)
public void interactWithKeyboard()
Your project should only use standard java libraries (imported from java.*) or any libraries we provided with your repo. Your final submission for the Phase 2 Autograder and Checkoff should not use any external libraries other than the ones provided in the skeleton.
IMPORTANT NOTE: Do NOT use static variables unless they have the final keyword! In 2018, many students ran into major debugging issues by trying to use static variables. Static non-final variables add a huge amount of complexity to a system. Additionally, do not call
in as this will cause the autograder to exit and fail.
Phase 1: World Generation
As mentioned above, the first goal of the project will be to write a world generator. The requirements for your world are listed below:
The world must be a 2D grid, drawn using our tile engine. The tile engine is described in lab12.
The world must be pseudo-randomly generated. Pseudo-randomness is discussed in lab 12.
The generated world must include distinct rooms and hallways, though it may also include outdoor spaces.
At least some rooms should be rectangular, though you may support other shapes as well.
Your world generator must be capable of generating hallways that include turns (or equivalently, straight
System.exit()
interactWithInputString
hallways that intersect).
The world should contain a random number of rooms and hallways.
The locations of the rooms and hallways should be random.
The width and height of rooms should be random.
Hallways should have a width of 1 or 2 tiles and a random length.
Rooms and hallways must have walls that are visually distinct from floors. Walls and floors should be visually distinct from unused spaces.
Rooms and hallways should be connected, i.e. there should not be gaps in the floor between adjacent rooms or hallways.
All rooms should be reachable, i.e. there should be no rooms with no way to enter
The world should be substantially different each time, i.e. you should not have the same basic layout with easily predictable features
As an example of a world that meets all of these requirements (click for higher resolution), see the image below. In this image, # represents a wall tile, a dot represents a floor tile, and there is also one golden colored wall segment that represents a locked door. All unused spaces are left blank.
Once you’ve completed lab 12, you can start working on your world generation algorithm.
It is very likely that you will end up throwing away your first world generation algorithm. This is normal! In real world systems, it is common to build several completely new versions before getting something you’re happy with. The room generation algorithm above was my 3rd one, and was ultimately much simpler than either of my first two.
You’re welcome to search the web for cool world generation algorithms. You should not copy and paste code from existing games or graphical demos online, but you’re welcome to draw inspiration from code on the web. Make sure to cite your sources using @source tags. For inspiration, you can try playing existing 2D tile based games. Brogue is an example of a particularly elegant, beautiful game. Dwarf Fortress is an example of an incredibly byzantine, absurdly complex world generation engine.
Tileset and Tile Rendering
The tile rendering engine we provide takes in a 2D array of
objects and draws it to the screen. Let’s call this for now. corresponds to
TETile[][] world
the bottom left tile of the world. The first coordinate is the x
world[0][0]
coordinate, e.g. refers to the tile 9 spaces over to the right from the bottom left tile. The second coordinate is the y coordinate, and the value increases as we move upwards, e.g. is 5 tiles up from the bottom left tile. All values should be non-null, i.e. make sure to fill them all in before calling . Make sure you understand the orientation of the world grid! If you’re unsure, write short sample programs that draw to the grid to deepen your understanding. If you mix up x vs. y or up vs. down, you’re going to have an incredibly confusing time debugging.
We have provided a small set of default tiles in
and these should serve as a good example
of how to create objects. We strongly recommend adding your own tiles as well.
The tile engine also supports graphical tiles! To use graphical tiles, simply provide the filename of the tile as the fifth argument to the constructor. Images must be 16 x 16, and should ideally be in PNG format. There is a large number of open source tile-sets available online for tile based games. Feel free to use these.
Any object you create should be given a unique character that other tiles do not use. Even if you are using your own images for rendering the tile, each should still have its own character representation.
If you do not supply a filename, or the file cannot be opened, then the tile engine will use the unicode character provided instead. This means that if someone else does not have the image file locally in the same location you specified, your world will still be displayed, but using unicode characters instead of textures you chose.
Tileset.java
world[0][5]
renderFrame
world[9][0]
The tile rendering engine relies on . We recommend against using commands like
or unless you really know what you’re doing, as you may considerably alter or damage the a e s t h e t i c of the system otherwise.
Starting Your Program
Your program will be started by running the method of the class. You will see that this method calls one of two possible methods, based on the way a user wants to interact with your program. Users can provide the program a command line argument, describing how they want to generate the random world and what exploration they wish to complete, in which case, calls the
When running with an input string, the format of the command argument should be
, where is the input of
. If no command line argument
was provided, the user wants to interact with the program using a GUI menu and exploring the world using the keyboard, in which case calls the
For phase 1, your project does not need to support but it must support
. Specifically, you should be able to handle an input of the format where
each # is a digit and there can be an arbitrary number of # s. This corresponds to requesting a new world ( N ),
providing a seed ( # s), and then pressing S to indicate that the seed has been completely entered.
Core.Engine.interactWithInputSting(String s)
-s inputString
interactWithInputString()
Core.Main.main
inputString
Core.Engine.interactWithKeyboard()
interactWithKeyboard()
interactWithInputString()
“N#######S”
We recommend that you do not implement
until you get to phase 2 of the project ( interactivity), though you’re welcome
to do so at anytime. It will be easier to test drive and debug your world generator by using
We also recommend not using until you get to phase 2. It’s easier to do phase 1 by writing your own
method in the class you interact with the most, that creates a renderer, calls , then draws the result. If you write your own method, you can ignore the -s flag and make the input string your only program argument. To provide a method command line argument through IntelliJ, if you’ve already tried to run the method, go to Run > Edit Configurations, and you can edit the Program Arguments box to input your string, as shown below.
Core.Engine.interactWithKeyboard()
interactWithInputString
Core.Main.main
interactWithInputString
程序代写 CS代考 加QQ: 749389476
If you haven’t run the method yet, right-click the green arrow next to your main method like shown below, and click “Modify Run Configurations” to access the same window as above.
Finally, we recommend that you make minimal modifications to the class. It is a much better idea to delegate all the work of the program to other classes you will create.
When your method is run, your program must display a Main Menu that provides at LEAST the options to start a new world, load a previously saved world, and quit. The Main Menu should be fully navigable via the keyboard, using N for “new world”, L for “load world”, and Q for quit. You may include additional options or methods of navigation if you so choose.
Core.Engine.interactWithKeyboard()
After pressing N on the keyboard for “new world”, the user should be prompted to enter a “random seed”, which is a long value of their choosing. This long data type will be used to generate the world randomly (as described later and in lab 12). The UI should show the seed value that the user has entered so far. After the user has pressed the final number in their seed, they should press S to tell the system that they’ve entered the entire seed that they want. Your world generator should be able to handle any positive seed up to 9,223,372,036,854,775,807. There is no defined behavior for seeds larger than this.
The behavior of the “Load” command is described later in this specification.
If the system is instead started with
, no menu should be displayed and nothing should be drawn to the
screen. The system should otherwise process the given String as if a human user was pressing the given keys using the method. For example, if we call
, your program should generate a world with seed 3412 and return
the generated 2D tile array. Note that letters in the input string can be upper or lower case and your engine should be able to accept either keypress (i.e. “N” and “n” should both initiate the process of world generation). You should ** NOT** render any tiles or play any sound when using .
If you want to allow the user to have additional options, e.g. the ability to pick attributes of their character, specify world generation parameters, etc., you should create additional options. For example, you might add a fourth option “S” to the main menu for “select creature and create new world” if you want the user to be able to pick what sort of creature to play as. These additional options may have arbitrary behavior of your choosing, however, the behavior of N, L, and Q must be exactly as described in the spec!
Requirements
For Phase 1, you should be able to run by providing an input String, and have your program create a world, that adhere to the requirements mentioned above
Core.Engine.interactWithInputString()
Core.Engine.interactWithKeyboard()
Core.Engine.interactWithInputString(“N3412S”)
interactWithInputString()
along with our randomness requirements mentioned in the Submission and Grading section below. Note that you should render the world to check your code by writing your own method, but for the autograder,
should not render the world, only returning the row as a array. Worlds should be
visibly different for different seeds provided to the program.
Design Document
Since we did not provide you with any significant skeleton code for Project 3, and since the project is very open-ended, we expect that BYOW implementations will vary a fair amount between students. We recommend that you have a design document that reflects the current state of your project.
Before you begin writing any code, use the guidelines listed here to create a plan for every feature of your BYOW program, and convince yourself that your design is correct. Writing a design document is an iterative process. After coming up with your initial design, you may find some flaws in it, requiring you to revisit your design and update its description according to your new findings.
Note: Your design document will NOT be graded. Design Document Guidelines
You may use the following format for your BYOW design document.
interactWithInputString()
Code Help
Design Document Sections 1. Classes and Data Structures
Include here any class definitions. For each class, list the instance variables (if any). Include a brief description of each variable and its purpose in the class.
2. Algorithms
This is where you describe how your code works. For each class, include a high-level description of the methods in that class. That is, do not include a line-by-line breakdown of your code, but something you would write in a javadoc comment above a method, including any edge cases you are accounting for.
3. Persistence
You should only tackle this section after you are done with Phase 1. This section should describe how you are going to save the state of a world, and load it again, following the requirements in the spec. Again, try to keep your explanations clear and short. Include all the components your program interacts with – classes, specific methods, and files you may create.
Phase 2: Interactivity
In the second phase of the project, you’ll add the ability for the user to actually interact with the world, and will also add user interface (UI) elements to your world to make it feel more immersive and informative.
The requirements for interactivity are as follows:
The user must be able to control some sort of “avatar” that can moved around using the W, A, S, and D keys. By “avatar”, we just mean some sort of on-screen representation controlled by the user. For example, in my project, I used an that could be moved around.
The avatar must be able to interact with the world in some way.
Your system must be deterministic in that the same sequence of key-presses from the same seed must result in exactly the same behavior every time. Note that a object is guaranteed to output the same random numbers every time.
In order to support saving and loading, your program will need to create some files in your directory (more details later in the spec and in the skeleton code). The only files you may create must have the suffix “.txt” (for example “save-file.txt”). You will get autograder issues if you do not do this.
Optionally, you may also include game mechanics that allow the user to win or lose. Aside from these feature requirements, there will be a few technical requirements for your system, described in more detail below.
UI (User Interface) Appearance
After the user has entered a seed and pressed S, the world should be displayed with a user interface. The user interface of your project must include:
A 2D grid of tiles showing the current state of the world.
A “Heads Up Display” (HUD) that provides additional information that maybe useful to the user. At the bare
minimum, this should include Text that describes the tile currently under the mouse pointer.
As an example of the bare minimum, the simple interface below displays a grid of tiles and a HUD that displays the description of the tile under the mouse pointer (click image for higher resolution):
You may include additional features if you choose. In the example below (click image for higher resolution), as with the previous example, the mouse cursor is currently over a wall, so the HUD displays the text “wall” in the top right. However, this HUD also provides the user with 5 hearts representing the avatar’s “health”. Note that this world does not meet the requirements of the spec above, as it is a large erratic cavernous space, as opposed to rooms connected by hallways.
As an example, the game below (click image for higher resolution) uses the GUI to list additional valid key presses, and provides more verbose information when the user mouses-over a tile (“You see grass-like fungus.”). The image shown below is a professional game, so we do not expect your project to have this level of detail (but we encourage you to try for some interesting visuals).
For information about how to specify the location of the HUD, see the
method of or see lab 12.
UI Behavior
After the world has been generated, the user must be in control of some sort of avatar that is displayed in the world. The user must be able to move up, left, down, and right using the W , A , S , and D keys, respectively. These keys may also do additional things, e.g. pushing objects. You may include additional keys in your engine. The avatar should not move when attempting to move into a wall and the program should not error.
The system must behave pseudo-randomly. That is, given a certain seed, the same set of key presses must yield the exact same results!
In addition to movement keys, if the user enters :Q (note the colon), the program should quit and save. The
initialize(int width, int height, int xOffset, int yOffset)
TERenderer
description of the saving (and loading) function is described in the next section. This command must immediately quit and save, and should require no further key-presses to complete, e.g. do not ask them if they are sure before quitting. We will call this single action of quitting and saving at the same time “quit/saving”. This command is not case- sensitive, so :q should work as well. Additionally, : followed by any other letter should not do anything.
This project uses to handle user input. This results in a couple of important limitations:
does not support key combinations. When we say :Q , we mean : followed by Q .
It can only register key presses that result in a char. This means any unicode character will be fine but keys such as the arrow keys and escape will not work.
On some computers, it may not support holding down of keys without some significant modifications; i.e. you can’t hold down the e key and keep moving east. If you can figure out how to support holding down of keys in a way that is compatible with , you’re welcome to do so.
Because of the requirement that your system must handle String input (via ), your engine cannot make use of real time, i.e. your system cannot have any mechanic which depends on a certain amount of time passing in real life, since that would not be captured in an input string and would not lead to deterministic behavior when using that string vs. providing input with the keyboard. Keeping track of the number of turns that have elapsed is a perfectly reasonable mechanic, and might be an interesting thing to include in your world, e.g. maybe the world grows steadily darker with each step. You’re welcome to include
interactWithInputString
interactWithInputString
other key presses like allowing the user to press space bar in order to wait one turn.
If you’re having trouble getting started on how to implement user interaction, check out for inspiration.
Saving and Loading
Sometimes, you’ll be exploring your world, and you suddenly notice that it’s time to go to watch a CS 61B lecture. For times like these, being able to save your progress and load it later, is very handy. Your system must have the ability to save the state of the world while exploring, as well as to subsequently load the world into the exact state it was in when last saved.
Within a running Java program, we use variables to store and load values. Keep in mind that when your program ends, all the variables will go out of scope. Thus, you will need to persist the state of your program on some files that your program should create.
When the user restarts and presses L , the world should be in exactly the same state as it was before the project was terminated. This state includes the state of the random number generator! More on this in the next section. In the case that a user attempts to load but there is no previous save, your system should simply quit and the UI interface should close with no errors produced.
In the base requirements, the command :Q should save and completely terminate the program. This means an input string that contains :Q should not have any more characters after it and loading a world would require the program to be run again with an input string starting with L .
InputDemo.java
byow.Core.Main
Interacting With Input Strings
must be able to handle input strings that include movement
For example, the string corresponds to the user creating a world with the seed 543, then moving up four times, then left twice. If we called
, your system would return a representing the
world EXACTLY as it would be if we’d used
and typed these keys in manually.
Since the system must be deterministic given a seed and a string of inputs, this will allow users to replay exactly what happened for a given sequence of inputs. This will also be handy for testing out your code, as well as for our autograder.
must also be able to handle saving and loading in a replay string, e.g. would correspond to starting a new world
with seed 25, then moving right, right, up, right, then quit/saving. The method would then return the 2D
array at the time of save. If we then started the engine with the replay string , we’d reload the world we just saved, move right four times, then return the 2D
array after the fourth move.
Your world should not change in any way between saves, i.e. the same exact should be returned by the last call to for all the following scenarios:
Core.Engine.interactWithInputString(String s)
interactWithKeyboard
N543SWWWWAA
Core.Engine.interactWithInputString(“N543SWWWWAA”)
Core.Engine.interactWithInputString(String s)
TETile[][]
N25SDDWD:Q
TETile[][]
TETile[][]
TETile[][]
interactWithInputString
interactWithInputString(N999SDDDWWWDDD)
interactWithInputString(N999SDDD:Q)
interactWithInputString(LWWWDDD)
interactWithInputString(N999SDDD:Q)
, then , then
, then , then
we then called with input , we’d expect the exact same world state to be saved and returned as as with the previous call where we provided .
You do not need to worry about replay strings that contain multiple saves, i.e. is not considered a valid replay string, since the program should have terminated before the second :Q . You do not need to worry about invalid replay strings, i.e. you can assume that every replay string provided by the autograder starts with either or
L , where # represents the user entered seed.
The return value of the method should not depend on whether the input string ends with :Q or not. The only difference is whether the world state is saved or not as a side effect of the method.