Why Discuss Technical Design Before Implementation?
Overview
This learning module helps junior developers understand why creating a technical design PR before implementation leads to better outcomes, faster delivery, and reduced rework.
Instructions
Present the following content interactively. After each section, use AskUserQuestion to offer 3 options:
- "Continue to next section"
- "Show me an example"
- "I have a question"
Section 1: The Purpose of Technical Design PRs
Explain that in story-flow, before writing implementation code, developers create a pull request containing their technical design:
A technical design PR serves three purposes:
- Early feedback on approach - Get team input before investing time in implementation
- Alignment before coding - Ensure everyone agrees on the direction
- Living documentation - The merged design becomes part of the repo history
The key insight:
A technical design PR is not about seeking permission. It's about leveraging collective expertise to find the best solution before committing to an approach.
What gets included:
- Story requirements (from the story tracker, Figma analysis)
- Planned tasks with execution order
- Pseudo code, example payloads, or expected API responses
- Questions or concerns for reviewers
Section 2: The Shift-Left Concept
Explain the "shift-left" principle:
Shift-left means moving quality activities earlier in the development lifecycle.
Traditional approach (problems found late):
Design → Code → Code Review → QA → Production
↑
Problems found here
(expensive to fix)
Shift-left approach (problems found early):
Design Review → Code → Code Review → QA → Production
↑
Problems found here
(cheap to fix)
Why "left"?
If you imagine the development timeline flowing left-to-right, shifting activities to the left means doing them earlier.
The cost curve:
| Stage Where Issue Found | Relative Cost to Fix | |------------------------|---------------------| | Design review | 1x | | Code review | 5x | | QA testing | 10x | | Production | 100x |
Issues found later require:
- Undoing completed work
- Retesting affected areas
- Potential customer impact
- Emergency context-switching
Section 3: Cost-Effectiveness Benefits
Present the concrete benefits:
1. Avoid Wasted Implementation Effort
Without design review:
Developer spends 3 days implementing Feature X
→ Code review reveals a simpler approach exists
→ 3 days of work discarded
→ Rework begins
With design review:
Developer spends 2 hours writing design PR
→ Team suggests simpler approach
→ 2 hours adjusted
→ Implementation proceeds correctly
2. Reduce Rework Cycles
| Approach | Typical Iterations | |----------|-------------------| | Code first, discuss later | 3-5 review cycles | | Design first, code second | 1-2 review cycles |
3. Faster Overall Delivery
Paradox: Adding a "planning step" actually speeds up delivery.
Without planning: [3 days coding] + [2 days rework] = 5 days
With planning: [2 hours design] + [2 days coding] = 2.25 days
4. Knowledge Sharing
Design PRs help the team:
- Understand upcoming changes
- Suggest improvements from past experience
- Identify conflicts with parallel work
- Onboard new team members through design history
Section 4: What to Include in a Technical Design PR
Present the recommended structure:
In the Story Markdown File
1. Story Requirements
- Acceptance criteria from the story tracker
- Insights from Figma designs
- Business context and user impact
2. Planned Tasks
- Ordered list of implementation steps
- Dependencies between tasks
- Which tasks can run in parallel
- Each task includes technical approach details:
- Key design decisions
- Pseudo code for complex logic
- Example payloads or API contracts
// Example task with embedded technical approach:
// Task: Add user authentication endpoint @src/api/auth.ts
// - Use JWT tokens with 24h expiry
// - Store refresh tokens in httpOnly cookies
//
// Expected API response:
interface LoginResponse {
accessToken: string;
expiresIn: number;
user: {
id: string;
email: string;
};
}
In the Pull Request
3. Alternatives Considered (include in PR description or comments)
| Approach | Pros | Cons | Decision | |----------|------|------|----------| | Option A | Fast to implement | Hard to extend | | | Option B | Flexible | More complex | Chosen |
4. Questions for Reviewers (include in PR description or comments)
- Highlight uncertainties
- Ask for input on specific decisions
- Flag areas where you need guidance
Section 5: Common Mistakes to Avoid
Present each mistake with correction:
Mistake 1: Skipping Design for "Simple" Changes
"This is just a small feature, I'll skip the design PR"
→ Feature has hidden complexity
→ Major rework needed after code review
Better approach: Start with a lightweight design. If it's truly simple, the design takes 15 minutes and confirms that.
Mistake 2: Not Enough Implementation Detail
# BAD - Too vague
## Tasks
- Add authentication
- Update UI
# GOOD - Actionable detail
## Tasks
1. Create AuthService class with login/logout methods
2. Add JWT token storage in localStorage
3. Create ProtectedRoute component wrapper
4. Update Header to show user email when logged in
Mistake 3: Treating Design as Final
"The design is approved, I can't change anything now"
Reality: Designs evolve during implementation. When you discover something that changes the approach:
- Update the story markdown
- Comment on the PR explaining the change
- Continue implementation
Mistake 4: Rushing to Merge Without Feedback
"No comments after 4 hours, I'll merge and start coding"
Better approach:
- Directly ping reviewers if urgent
- Wait for at least one approval
- Silence doesn't mean approval
Mistake 5: Creating Design After Starting Implementation
"I'll code this first, then write up what I did"
This defeats the purpose. Post-hoc documentation:
- Doesn't catch design issues early
- Doesn't benefit from team input
- Becomes a justification, not a discussion
Section 6: Quick Reference Checklist
Present this checklist for the developer to keep:
Before submitting a technical design PR, verify:
- [ ] Story requirements are clearly documented
- [ ] Tasks are specific and ordered
- [ ] Dependencies between tasks are identified
- [ ] Technical approach includes concrete examples
- [ ] Pseudo code or payload examples for complex logic
- [ ] Alternatives were considered and documented
- [ ] Questions for reviewers are highlighted
- [ ] Related ticket ID is mentioned
During review:
- [ ] Respond to all comments
- [ ] Update design based on feedback
- [ ] Resolve addressed conversations
- [ ] Wait for at least one approval
After merge:
- [ ] Implementation follows the approved design
- [ ] Design updates are reflected in story markdown
- [ ] Major deviations are communicated to reviewers
Example Inputs
- Learn about technical design PRs
- Why should I create a design before coding?
- What is shift-left?
- How do I write a good technical design?
- Teach me about planning before implementation