Rubric for labs

This document lays out how we’ll be grading labs at a high level: what are we looking for in order for a submission to achieve a certain grade?

Some things to note:

this rubric is intended to be general; not all points below will apply to all labs. For example, lab 1 will not be graded on its big-O performance (since we have not gone over big O at the time of lab 1).
An “A” grade means that you submitted excellent work. An “A+” means that you went well above and beyond what was necessary. Consequently, expect A+ grades to be rare.

Functionality (50%)

Functionality is judged by running your program and inspecting your code. If you program does not produce the expected results, partial credit will be assigned based on the specific functionality specified in the lab assignment. A well-designed program will comprise several independent units that can each be tested and verified in isolation. Writing your programs this way will ensure that we can award maximum credit for the work that you complete.

Here is a breakdown of the functionality scale:

E: Lab was not attempted.

D: Lab is largely incomplete; few, if any, concrete tasks were identified or implemented.

C: Lab is largely incomplete, but concrete tasks have been identified, and one or more of those tasks have substantial implementation.

B: Most lab infrastructure is present, but a small number of key components are largely incorrect.

A: Functionality is mostly correct or close to correct, but some corner cases are missing.

A+: Functionality is not just perfect, it goes above and beyond by adding a degree of “polish” or additional features to the code.

Design (30%)

Design is judged based on some of the key principles we have discussed in class and in the readings. Your code should use abstraction in order to hide/isolate implementation details of classes and your code should be modular and reusable. This means that the classes that we write should be self-contained and have their functionality clearly defined.

Here is a breakdown of the design scale:

E: Lab was not attempted.

D: Code is disorganized and does not follow discernible design principles.

C: Code misses many key design principles:

Code uses reasonable class structure, but may not wisely allocate functionality or define clear abstraction boundaries.
Fields and methods may be designated as public, private, or protected haphazardly.
Functions may be too long or may be unclear.
Global variables may be used to communicate across methods, rather than performing appropriate parameter passing.

B: Code largely applies design principles, but falls short in more than one key area.

Inconsistent/illogical application of public/private/protected visibility modifiers for effective information hiding.
Lacks well-defined class boundaries.
Large tasks could be better decomposed into concrete sub-tasks as (helper) methods.
Unnecessary variables and/or extraneous code.
Methods should only communicate using parameters and return values, rather than making assumptions about the state of global variables.

A: Code follows/practices good design principles, but there are one or more clear areas for improvement, including but not limited to:

Aspects of code could be further simplified by restructuring methods and/or by utilizing existing interfaces.
Classes or methods could be consolidated by removing unnecessary/redundant variables.
Program is inefficient with respect to Big-O.

A+: Code is exceptionally elegant, efficient, and clear. This grade usually results when a professor says “wow, I wish I wrote code like that!”

Documentation/Style (20%)

Documentation/style are judged on several components:

Class-level comments (mandatory)
Method-level comments (mandatory)
Member variables (documented as needed)
In-line comments (describe non-intuitive code as needed)
“Overall Quality” of comments/variable naming
checkstyle error compliance
Other style guidelines as described in a particular lab

A score is first calculated using on the rubric below, then then 1 point is deducted for each category of failed checkstyle rule or lab-specific style guideline violation.

Here is a breakdown of the documentation/style scale:

E: Code is not documented.

D: Code is poorly documented:

Comments are sparse and/or insufficient to understand the code, alternatively, comments may be unhelpful (e.g., comment on what rather than why).
Variable/method names are not descriptive.
Code is not formatted for readability.

C: Code documents many methods and classes, but is not thorough:

Comments may describe obvious things like //member variables, rather than details about how the variables should be interpreted/used.
Comments exist, but are missing components (e.g., pre/post conditions for public methods)
Variable/method names may not always be well-chosen or descriptive.

B: Application of documentation and style largely meets the requirements for a score of 4 or higher, but is inconsistent. One or more classes/files are undocumented or poorly documented.

A: Documentation is present and useful:

Variable/method names are usually descriptive and follow naming conventions.
Pre/post conditions are present for public methods and and assertions are used appropriately.
Code is formatted for readability (indentation and whitespace).
In-line comments document most non-intuitive code snippets.

A+: Clear, easy-to-read documentation and formatting:

Code comments follow Javadoc standards.
In-line comments document non-intuitive code snippets.
All variable/method names are descriptive and follow naming conventions.
Code is formatted for readability (indentation and whitespace) and is therefore easy to read and understand. Logical blocks of code are clearly delineated with whitespace.
Pre/post conditions are present for all public methods and and assertions are used appropriately as documentation and debugging.