#7: Trust the progress!

1. Adidas iOS App Architecture

This is a tech talk about the architecture of the Adidas iOS app. Some of the techniques that they’re using:

• Modular architecture
• SwiftUI
• TCA
• Tuist

It’s quite interesting that Adidas is using TCA and Tuist in their app. There a lot of arguments about TCA in the community, so it’s good to see a big company like Adidas using them.

2. Xcode Cache: Boost your iOS Testing and Deploy on GitHub Actions

The author mentioned a workflow to boost your CI/CD workflows on GitHub Actions by caching the Xcode cache. The steps are:

• Cache DerivedData/SPM resolved
• Delete old cache, Update cache on post merge only to prevent increase storage cache
• Get cache to use for PR

3. The Problem with withCheckedThrowingContinuation and Task Cancelation: A Video Capturing Example

The author state a problem when using withCheckedThrowingContinuation that can easily cause wasting resources. The root cause is the even if the Task is cancelled, the withCheckedThrowingContinuation block will still be executed.

The solution to resolve the issue is to use withTaskCancellationHandler(operation, onCancel), and perform the cleanup in the onCancel block (the block will be triggered when the Task is cancelled).

It’s an good article to read and take note 💪

4. The Definitive Guide to Technical Debt

This is a really detail article about Technical Debt. Since it’s a long article, I will summarize the key points here:

Definition:
Technical debt is something that doesn’t match with the business reality and needs.

Types of tech debt:
Sometimes the requirement changes (especially when your product move fast) → the current software stops working → tech debt. We have 2 kinds of tech debt:

• Tech debt from the start: Wrong from the start cuz we can’t adapt the business requirement ⇒ Bad debt
• Tech debt over time: due to requirement change.

How to prevent:

• Understand the requirements before start coding:
  • Why need this requirement, is it finalized, do we need to support scalable, …
• Good design: Clean and scalable design (but don’t over-engineer)
• Minimize the chance of tech debt:
  • Remove unused code
  • Add tests
  • Refactoring code that’s not scalable/maintainalbe in long term.

How to solve

• Small: Do it right away since it small and quick to fix
• Medium: The whole team spend a dedicated day per week on the tech debt, rotation assigning to fix, …
• Large: Analyze, investigate and plan

Measure tech debt:

• Impact: business metrics impacted by the debt, or the value you unlock
• Fix cost
• Contagion: The cost to fix will higher overtime → should prioritize, …

If you want to learn more about how to deal with legacy code, you can read my book review here

5. How to Create Diagrams in Draw.io Using ChatGPT

The author shows how to use ChatGPT & Mermaid.js code to generate diagrams in Draw.io. It’s a good way to generate diagrams quickly and easily. Great article!

6. How I Ship Projects at Big Tech Companies

The question that I faced recently is: “Should you do what you think is right or what the boss wants you to do?”
Sometime the bosses doesn’t care about the quality of the codebase, and they only care about delivery more features. As an engineer, we know that rushing features will lead to tech debt, and it will slow down the development in the long run.

In the article, Ryan mentioned some key points:

• We should focusing on user-impact oriented: easy to market your work, level up your level, and a good engineer should know how to push back the management to avoid work that isn’t impactful
• “Although companies get more political as they grow, even the biggest companies (e.g. FAANG) have designed their organizations to minimize this. These companies are data-driven where possible and reward people who improve metrics that help the business.” - This is a nice idea to minimize the political in big tech companies and maintain good work culture.

If your organization values pleasing leadership over impact, I’d recommend switching teams. Your career will grow faster in the long term if you go where people are prioritizing impact. Not to mention that doing work that matters is much more fulfilling than pleasing people who may be wrong.