Overview and Grading
Overview and Grading
This page is the one place to look for what your team hands in, and how it is graded. The point values below come from the course’s Gradescope assignments; your course’s Gradescope is the authority for due dates and for any value that differs from this page.
The project at a glance
- You work in a team of 4 on one shared GitHub repo, with the app deployed throughout; the deployed app is what gets graded.
- Iteration 0 is setup. It has no deliverable of its own, but Iteration 1 cannot be graded until it is done: a stranger can log in to your deployed app.
- Iterations 1, 2 and 3 each follow the same rhythm: a planning meeting at the start, stories as GitHub Issues, branch-per-feature with pull requests and CI, a retrospective at the end.
- Every iteration is worth 50 points: 35 for the team (A, B and C below) and 15 for you individually (D). The weights inside the team deliverable shift as the project goes on, from set tasks in Iterations 1 and 2 to your own features in Iteration 3.
The four deliverables of every iteration
A. Planning Meeting (team, online form, early in the iteration). Your team number and meeting time; which members (if any) were absent without warning; who is pairing with whom. A checklist of what came out of the meeting: the stories you created as GitHub Issues from the iteration’s spec (with links), the points you assigned to them and who owns each, and when you will meet again to check progress. A link to (or upload of) the meeting’s notes. In Iteration 3 the planning submission is where your team proposes its two features, and staff agree to them.
B. Code Submission (team, Gradescope programming assignment). Your repository, submitted through Gradescope’s GitHub connection at the end of the iteration. It is graded by hand on whether all of the iteration’s code is there; there is no autograder.
C. Video, Links, and Assessment (team, online form). The bulk of the team grade. Video: a short video (2-4 minutes in Iteration 1, 2-5 minutes after that) in which every team member takes part, walking through the deployed app with its URL visible, demonstrating every feature you built and pointing out bugs and unfinished work. Feature work: graded task by task from the video and the deployed app, with the pull requests as evidence, and in Iterations 2 and 3 a check that the earlier iterations still work, login included. In Iteration 3 this is your two features, each graded on working end to end and on explaining what it does and why ActionMap needs it, plus a technical checklist that the two features must cover between them. Team work: merged pull requests reviewed by a teammate, the iteration’s issues with who did what and what was delivered, and fully passing CI runs. Testing: a coverage report with a plan for what to cover next, the accessibility scenarios you added, and two test cases you wrote with an explanation of what each one checks and why. Generative AI: how the team used AI this iteration, with a chat linked; in Iteration 3 this is a larger part of the grade, feature by feature.
D. Retrospective (individual, online form). Hold the retrospective as a team (the form explains how: a shared retro board, topics, three votes each, discussion in order of votes, action items). Then each member submits a link to or screenshots of the retro board and its action items, the meeting’s logistics, a few facts about your own iteration (who you paired with, how many standups, hours spent, a link to your commits), any private feedback for course staff, and confirms the two Teammates surveys (peer review and self-assessment), which are checked against Teammates. The rest is about your own work. GitHub contributions: links to every pull request you authored, every code review you gave (more than “LGTM”), and every issue you created, edited or commented on, with one of each explained. Assign your PRs, issues and reviews to yourself and to any teammates who worked on them with you: the assignees are how we credit them. AI chats: links to (or transcripts of) the AI chats you had while working, and which one was most useful.
Points per iteration
| Deliverable | Iteration 1 | Iteration 2 | Iteration 3 |
|---|---|---|---|
| A. Planning Meeting (Team) | 4 | 4 | 4 |
| B. Code Submission (Team) | 1 | 1 | 1 |
| C. Video + Links Submission (Team) | 30 | 30 | 30 |
| D. Retrospective (Individual) | 15 | 15 | 15 |
| Team total | 35 | 35 | 35 |
| Individual total | 15 | 15 | 15 |
| Total | 50 | 50 | 50 |
The team total is what the team earns together (A, B and C); the individual total is your own retrospective (D).
How each deliverable is graded
Planning meeting (4 points). Meeting information, 1 point. The checklist, 2 points: full credit for stories created as useful GitHub Issues with links (in Iteration 3, two feature proposals, each with a user and a reason, a sketch of how it works, the technical checklist items it covers, and issues with owners); 1 point when there is only one issue or feature, or they are bare. Meeting notes, 1 point: half if they are very bare, 0 if missing, blank or not accessible to staff.
Code submission (1 point). Full credit when all the iteration’s code is there, submitted from the team repo; half when it shows little evidence of progress; 0 for little or no code.
Video (2 points). A 2-4 minute (Iteration 1) or 2-5 minute (Iterations 2 and 3) video with all participating members, walking through the deployed app. Minus 1 for running over 5 minutes or under 2, and minus 1 if the deployed URL is not visible or the demo runs on localhost. The feature work shown in the video is graded separately, below.
Feature work (12, 12, 8 + 6 points). Iteration 1: 3 points for each of Tasks 1.1, 1.2 and 1.3 (3 complete, 2 with minor bugs, 1 attempted), and 3 points for Google or GitHub login working on the deployed app for someone outside the team (2 when it only works for the team’s own accounts, 1 when it only works locally). Iteration 2: 2 points for the Issues field (Task 2.1), 2 points for each of Tasks 2.2 to 2.5 of the option you chose (minus 1 per task with minor bugs, minus 2 per task that is incomplete), and 2 points for Iteration 1 still working on the deployed app, login included. Iteration 3: 4 points per feature, 3 for working end to end on the deployed app (2 with minor bugs, 1 attempted) and 1 for an explanation that names the user and the need it meets; and a 6-point technical checklist that the two features cover between them: a schema change beyond adding columns (2), new routes and controller actions (1), new views (1), and JavaScript your team wrote that enhances a view (2).
Team work (6, 6, 3 points). Iterations 1 and 2: 2 points for two merged pull requests each reviewed with comments by a different teammate (1 when they are self-merged or the reviews are only “LGTM”); 2 points for the iteration’s issues, each with an assignee and a reported status of delivered, finished or not started; 2 points for two fully passing runs of the All Specs workflow (1 each; minus 1 for linter failures; a run that only passes the linters does not count). Iteration 3: 1 point for issues and reviewed PRs, and 2 points for a green run on main after the final merge with Iterations 1 and 2 still working on the deployed app.
Testing (8, 8, 3 points). Coverage (4, 4, 2 points): full credit for 50% or more of app/ covered in Iteration 1 and 75% or more in Iterations 2 and 3, scaled down in tiers (Iteration 1: 3 for 40-50%, 2 for 30-40%, 1 for 20-30%; Iteration 2: 3 for 65-75%, 2 for 55-65%, 1 for 45-55%; Iteration 3: 1 for 65-75%). From Iteration 2, coverage of 80% or more earns 1 to 3 points of extra credit; in Iterations 1 and 2 a missing or vague coverage plan costs a point. Accessibility scenarios (2, 2, 1 points): 1 point for each scenario that is written and passing, minus half a point for one that does not pass, and minus 1 if the two scenarios overlap; in Iteration 3, 1 point for a passing scenario on each feature’s new page. Test cases (2 points in Iterations 1 and 2): full credit when both test cases are linked and each explanation names the behaviour under test and why it matters; half when only one is given or the explanations only restate the code.
Generative AI (2, 2, 8 points). Iterations 1 and 2: 2 points for naming the tools and what each person used them for, with one concrete example, chat linked, of correcting a model’s output or of a prompt that needed no changes. Iteration 3: 3 points per feature for the chats and concrete examples of what the team kept, modified and rejected, with the reasoning; 2 points for how the team’s use of AI changed over the project. A team that did not use AI and says so gets full credit.
Bonus (Iteration 3, up to 8 points). Up to 4 points for a Rails feature not covered in class (ActiveStorage, ActionText, ActionCable or Turbo Streams, a real job queue, caching, a mailer…) working in one of your features, and up to 4 for an additional service the deployed app depends on (Redis, object storage, mail; error monitoring counts for at most 1, with evidence it found something). Each needs a justification: why the feature needed it and what you learned. Bonus points count only when the required work has no significant defects.
Retrospective (15 points). Retro board (3 points): full credit when it is included, visible to staff, and shows topics, votes, discussion and action items; half when incomplete or inaccessible. Teammates surveys (2 points): 1 for the peer reviews and 1 for the self-assessment, checked against Teammates. GitHub contributions (8 points): 3 for the pull requests you authored, 3 for the code reviews you gave, 2 for the issues you created or shaped; full credit for substantive, assigned work with one of each explained, half when it is trivial, unreviewed, a bare “LGTM”, or not assigned to you. AI chats (2 points): full credit for chats shared with a reflection on what was useful, half for one without the other; a student who did not use AI and says so gets full credit.
Late submissions and adjustments
Every Gradescope assignment accepts late submissions until a hard deadline two or three days after the due date; check each assignment on Gradescope. Beyond the rubrics above, your TAs may adjust a team’s or an individual’s grade based on the peer evaluations, the quality of your stories, and how the team worked together in pull requests and issues. The goal is a healthy development process, not just a finished app.