As I reflect on my 24-year journey in software development—transitioning from developer to architect to product manager—it’s clear that my perspective has evolved significantly. Early on, I was a purist, firmly committed to doing things the “right way” and adhering to traditional architectural principles. However, the landscape of software development has transformed dramatically, especially with the advent of tools like Git, CI/CD pipelines, microservices, and modern UX frameworks such as React.js and Angular.

Despite having access to powerful methodologies and practices, I often sensed a hesitation to think creatively and deliver quickly. Today’s engineering landscape encourages innovation alongside risk management. My understanding of good architecture has shifted; it’s no longer simply about rigid adherence to established principles, but rather about effectively serving our customers.

By isolating functionality, I’ve realized that interim solutions—like quickly leveraging a Google Translate API for translations—can fulfill immediate needs without jeopardizing long-term goals. This flexible mindset allows us to replace these temporary fixes with polished, longer-term architectures over time. One defining moment in this evolution stemmed from a past experience where architects chose to rebuild a feature that already had 90% of the functionality. This decision, while aligned with ideal architecture, led to a staggering delay of over two and a half years, ultimately at the expense of our customers.

Another instance involved developing a proof of concept with the Google Translate API to promptly address UI translation issues. Though it was a practical solution for an urgent problem, the desire to utilize an existing comprehensive framework caused significant delays. With modern tools, isolating such functionality could have served as a strategic interim solution, allowing for incremental enhancement while delivering immediate value.

Today, I champion practicality and agility as much as I value solid design. The tools and architectures we possess today facilitate managing and replacing interim solutions without accruing substantial technical debt. They grant us the flexibility to prioritize customer needs without sacrificing long-term quality.

Modern frameworks like React and Angular have introduced new paradigms for software development, promoting decoupling and cohesion in codebases. With capabilities such as dependency injection, state management, and automated testing, developers can confidently implement incremental changes across systems. This evolution prompts a critical reflection: what truly defines “good code”?

The notion of good code may need redefinition as we navigate AI-generated solutions. If our codebases are organized into independent modules, backed by automated tests, then why should AI-generated code that adheres to these principles not be considered “good”? It’s time to recognize that good code is more about maintainability, adaptability, and cohesion than about strict adherence to outdated standards.

We must strike a balance between supporting foundational principles and being responsive to technological advancements and market needs. Our goal should be to create isolated, modular solutions that can evolve seamlessly, all while keeping the customer experience paramount. Ultimately, ‘doing it right’ involves a commitment to delivering value through design, timely execution, and responsiveness, reflecting the flexible mindset required to adapt and grow in an ever-changing tech landscape.