Why Old Tech Like SQL Still Beats New Hype – Lessons on TypeScript and Docker

Development is a rapidly evolving field, with new buzzwords from web and mobile to cloud, containers, DevOps, big data, front‑end frameworks, blockchain, and AI, while many developers chase the latest tools.

However, several seasoned engineers argue that some "old" technologies are undervalued, and that newer hype does not always deliver the expected productivity gains.

SQL Is the Most Valuable Skill in My Career

Original article: http://www.craigkerstiens.com/2019/02/12/sql-most-valuable-skill/

InfoQ translation: https://www.infoq.cn/article/Lg5quPMJ1lKXm-5o46xW

The author, a cloud‑service provider focused on PostgreSQL, claims:

SQL is the most valuable skill I have learned. It is valuable across many roles and disciplines. Once mastered, it requires no relearning. It makes you look like a superhero; those who know it appear especially powerful.

My view: I partly agree. Many NoSQL solutions initially claim to differ from traditional RDBMS, yet they eventually adopt SQL‑like features. SQL, however, lacks built‑in distributed execution concepts, which modern data platforms strive to address. Often, small‑scale data problems can be solved directly with SQL.

The TIOBE language trend report supports this: among the top 20 languages, only SQL has remained remarkably stable since 2004, showing its enduring reliability.

From a long‑term perspective, SQL may not promise the highest returns, but it offers consistent, trustworthy value for developers.

Why TypeScript May Not Suit Large Projects

Original article: https://medium.com/javascript-scene/the-typescript-tax-132ff4cb175b

InfoQ translation: https://www.infoq.cn/article/Bmx*2UO9VRMTSbSWw4qX

Although the author enjoys TypeScript and uses it at scale, the evaluation concludes that TypeScript’s benefits, costs, and shortcomings do not meet expectations for large projects.

I have a deeper understanding of TypeScript’s pros, costs, and limitations. I would not use it in another large project unless it improves dramatically.

In other words, TypeScript adds value but not as much as anticipated, and its adoption cost may outweigh the benefits.

The author’s assessment method—scoring tooling, API docs, refactoring, training, hiring—offers useful insight. He also notes that type safety’s impact on bug density is limited compared to practices like unit testing, code review, and TDD, which can reduce bugs by over 90%.

Supporters tout type safety, yet evidence of its effect on bug density is scarce. Combining TDD with design and code reviews yields far greater bug reductions than type safety alone.

I agree that solid testing and review processes often deliver more improvement than new language features.

Docker: A Risky Bet

The main argument is that betting everything on Docker is dangerous. An alternative is deploying applications as large binaries or uber‑jars (e.g., Java, Go, Clojure) and managing servers with traditional tools like Chef, Puppet, or Ansible, while still using Kubernetes for orchestration when needed.

Docker adds complexity and requires deep system‑admin expertise; many articles oversimplify its use, giving a false impression of ease in production.

Docker should be seen as an advanced optimization, not a default solution. Its adoption raises system complexity and demands specialized knowledge. Most Docker tutorials showcase trivial use cases, ignoring the challenges of multi‑host production environments. Blindly standardizing Docker from the start contradicts best practices that recommend incremental adoption after establishing stable systems.

Additional sources criticize Docker’s production reliability, highlighting file‑system and interface issues.

Conclusion

New technologies bring value, but we should not forget the reliable “old” tools that have stood the test of time. Understanding their failure modes often leads to lower operational overhead and more predictable outcomes.