How to Choose the Right Tech Stack: Efficiency, Ecosystem, and Team Dynamics

1 Preface

I have been a Java backend developer for seven to eight years, witnessing the evolution from jQuery and EasyUI to front‑end separation, from Hibernate to Spring MVC, Spring Boot, micro‑services, and finally to big‑data technologies like Spark, Flink, and various databases.

Throughout this journey I have questioned whether adopting a technology for its own sake is justified, especially when forced to use "outdated" tools.

2 Efficiency

2.1 No Absolute Efficiency

Comparing performance metrics without controlling variables is meaningless; a 5% speed gain may not outweigh other costs.

2.2 Is Efficiency Absolutely Important?

Even when benchmarks show ZeroMQ as the fastest message queue, most companies still choose RabbitMQ or ActiveMQ due to familiarity and ecosystem support.

3 Environment

3.1 Domestic Development Landscape

Tools like MyBatis dominate in China, while Hibernate remains popular abroad, as shown by Google Trends and StackOverflow tag counts.

Images illustrate the differing popularity trends.

The regional differences stem from factors like Alibaba's influence and hiring practices; developers tend to follow the dominant tools in their environment.

3.2 Impact of Technical Communities

English proficiency affects access to original documentation, but many Chinese developers rely on translated or second‑hand resources, leading to a closed learning loop.

Community preferences also shape tool adoption, e.g., Spring Cloud vs. Apache Dubbo.

4 Team

4.1 Leadership and Core Skills

A 2017 streaming project chose JStorm over Spark Streaming due to a leader’s preference, despite my familiarity with Spark. The decision led to later difficulties, highlighting that a leader’s lack of technical depth can increase operational risk.

Key takeaways:

Competitiveness comes from the ability to learn new technologies quickly and apply prior experience to solve problems.

Technology itself has no inherent superiority; choices often depend on team expertise and preferences.

Beyond learning cost, consider operational, recruitment, and maintenance costs when selecting a stack.

Author: 是春壹呀 (Source: cnblogs.com)