TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

The JavaScript “ecosystem” is a hot mess and so is software development

4 点作者 avrgamp超过 3 年前

2 条评论

genjipress超过 3 年前
&quot;That people find this level of ludicruous confusion and complexity acceptable is mind blowing. And, to think that there is a whole generation of developers to whom this is the baseline of writing software, who think that this is normal, is genuinely painful. That there are people who knowingly perpetuate such complexity in software development in general, is painful. That reminds me of a developer that I spoke to who only knew how to deploy a static website via a “CI&#x2F;CD” system connected to a K8s cluster. They did not know that it was possible to cp or rsync an index.html file to a directory on a Linux system running a web server. They were unable to even visualize that fundamental concept because CI&#x2F;CD and K8s was their baseline.&quot;<p>Slightly orthogonal to the point of the article, but I&#x27;ve had some firsthand experience with this particular insight. I was talking to someone who was deploying a complex application via CI&#x2F;CD to a cluster and was frustrated about some performance issue or other. This after much fribble about the pods and the cluster state and the load balancing and so on. I opined that maybe the problem was the application itself, since symptoms of that had persisted before the deployment, and asked if they had done any profiling. They said, &quot;What&#x27;s profiling?&quot;
winternett超过 3 年前
20+ years into my experience as an IT architect I&#x27;ve battled with overly complex ideas non-stop. The worst part of jobs is usually having to go through a huge research and documentation process just to get a customer to adopt the solution I&#x27;m proposing because of all of the frameworks and tools floating around with deeply false promises.<p>I&#x27;ve found it&#x27;s simply better to avoid working with devs that implement overly-complex and under-supported solutions.<p>A JS library, no matter how cool it works, often isn&#x27;t supported properly after the 1-2 year mark if it&#x27;s not maintained and updated responsibly. Less is more if a site or application runs flawlessly and if weak workarounds don&#x27;t constantly need to be involved in every product release.<p>Many devs create complex code because they think perhaps that that is the way to retain job security, it simply doesn&#x27;t work over time for anything beyond a portfolio site. In the business world, well-serving-mission-critical systems are either maintained by big teams and a lot of funding, or by tight knit teams using technology that is simplified and low-maintenance.<p>Blockchain, abstract JS frameworks, and many other new concepts are promoted these days as cutting edge, with buzz-word branding like &quot;Web 3.0&quot; etc... Because it&#x27;s very hard to disparage a history of stability and reliability found in PHP, Python, C, and Java (etc) without trying to rebrand stable and proven solutions as &quot;Old School&quot; and &quot;Legacy&quot;. Using those terms for proven solutions is wrong in my view, and it only serves to create a new revenue stream in refactoring and &quot;screwing up&quot; an ecosystem that&#x27;s worked. Those &quot;Legacy&quot; ecosystems are still being updated by their support communities, and likely by leveraging far more organized, tested, and stable methods than those used by creators of newer frameworks.<p>In contrast, it&#x27;s very hard, expensive, and time consuming to do a fully functional proof of concept and to maintain it over long periods of time especially when your solution is based on a small and fragile community of support, and with very little documentation because of that solution&#x27;s infancy.<p>Stable (for me) always better than cutting edge in every solution I&#x27;ve built. The tools are not as important as the end product, it&#x27;s fulfillment of business objectives, reliability, uptime, ease of use, and it&#x27;s functional precision.<p>Design and Innovation do still have great merit in solutions, but they should never outweigh nor outrank reliability, efficiency (simplicity), consistency, and function of a solution.