Thinking about the end of it at @legacynotes_ch, in a long-term relationship with @McDonalds, bringing techy and designey people together at @frontzurich.
If I had to pin it down to only one thing, it would be culture. There is a perfect storm of several factors that has caused the rapid descent over the last 10 or so years. Here are a few recurring themes I've seen as a consultant, trainer, and engineering leader during that time:
1. Maturity - We have an overwhelmingly junior-heavy JS community, and no clear mentorship path for these folks. This leads to over-dependence on frameworks and unhealthy hero worship of social media tech influencers. It also means that innovation is only happening within a small group and most people don't know how to solve problems and create solutions for themselves.
2. Knowledge - There is a broad lack in understanding of fundamental Web Standards, even among many Sr+ engineers. There is also a broad ignorance of the history of web tech and software engineering more broadly. This leads to poor technical decision making, both at the architecture level and the implementation level. This has also led to the adoption of overly complicated solutions to relatively simple problems.
3. Hubris. - While there has been so much talk of imposter syndrome in the industry, ironically, at least on social media, folks seem to be overly confident in their own skills, overly confident in their chosen frameworks, and overly confident in their social media celebrities.
4. Tribalism - Many individuals in our industry associate themselves with a sort of "tribe" surrounding a particular library/framework or social media influencer. This can make it nearly impossible to teach individuals new things, difficult for them to cope with changes happening in the web platform that "threaten" what they have built their career or personal brand around, and nearly impossible for them to handle criticism. Objective evaluation of alternative technologies is right out.
If we could address the culture issues, then the technical challenges would solve themselves.
@brianleroux@ryanflorence I’m starting to warm up to the idea of having only a form component rather than form and fields. This way, there is no shadow boundary between them and a field without a wrapping form is not very useful anyway.
@DanielWirth2 Frage zu https://t.co/dTbqgTECLP: Ist es nicht bereits heute so, dass Rollstuhlfahrer:innen nur "an einem Ort ohne Hilfe in die Züge rollen können"? Was wird denn in dieser Hinsicht genau verändert?
FUD:
Web components suck! No one is using them!
But is that true?
Google:
Our annual $35 billion revenue YouTube product is built on Web Components.
Microsoft:
We converted our entire content platform from React SSR to Web Components which doubled the perf. We finally crossed the $10 billion annual revenue milestone.
Adobe:
We built Photoshop Web with Web Components.
Salesforce:
Our $31 billion annual revenue platform has been based on Web Components for years.
SpaceX:
We built our dashboard system with Web Components and launched rockets into space.
GitHub:
We've been building parts of the GitHub experience with Web Components for years.
Anyone else building with Web Components?
Amazon, Apple, Alaska Airlines, BBVA, Begin, Blizzard, BYU, Clever Cloud, Cloud Four, Comcast, EA, 11ty, ESRI, Ford, GE, GitLab, General Motors, GuideCom, IBM, Infragsitics, ING, The Internet Archive, Ionic, Joomla!, NASA, Netlify, Penn State, Reddit, Red Hat, SAP, Scania, Stripe, Ubisoft, Vaadin, Visa, Wolkswagen, Wordle, etc.
https://t.co/uhb9mMgxJK
@kettanaito Doesn’t have to be. You can get very far with the node:http library. The platform does so much already, but many people seem to prefer the most complex abstractions over learning the basics. 🤷♂️
A simple server with basic routing, body parsing, and dev watch mode with less than a gigabyte of npm dependencies? Is it possible?
Why I ❤️ `import http from "node:http”`:
https://t.co/JKdc6yAez7
It's so rewarding seeing people get excited about this! View Transitions FTW!
It's now possible to ship app-like behavior (persistent UI, persistent UI state, persistent scroll, styled page transitions) without giving up Astro's fast, simple, every-page-is-HTML-by-default foundation.
Such an amazing long-term payoff from our early bet on Island Architecture, almost 2 years ago. And we've only just scratched the surface...
Coming soon in Astro 3.0, but available to try experimentally today in 2.10!