 The problem with modern frontend isn't that people pick React, it's that a huge chunk of the industry has abandoned user-centric values in the name of efficiency, delivering neither good UX nor acceptable DX in the process. Picking React in '23 is merely a symptom of a much deeper rot. 26 comments
 @loke efficiency as in hot swapable developers because we can't be bothered to put resources into staff retention @slightlyoff  @mensrea @slightlyoff Right. Then I agree since react pages are anything but efficient. 🙂 @loke @mensrea Sorry for being obtuse; lots of orgs have picked up React because they were told it's industry-standard, that they would be able to hire for it easily, etc. etc. It turns out that React on its own doesn't get you very far, and so there's a ton of time and money that needs to be spent either building infra to support some agglomeration of state tools + SSR configs + component libraries, or investment in learning metaframeworks like Next. The savings never materialise. @slightlyoff i think there is another aspect to this. the "modern way" with frameworks, language abstractions (typescript), build pipelines, et al. make people think they're doing "real programming". so they keep layering complexity to solve problems caused by complexity because they're really smart @loke  @mensrea @loke @slightlyoff This. React is the ultimate in inefficiency. The goal isn't to be efficient but to maximally disempower and make fungible the labor involved. @dalias @mensrea @loke @slightlyoff That's it! Project management styles and tech choices are both shaped by the need to make the work of programmers predictable and measurable. Huge costs are incurred, and seen as entirely worthwhile.  @loke @slightlyoff Sounds like profitability is all. The less money handed to developers and customers out of the current revenue base means more money handed to shareholders. @slightlyoff The company's ability to have a pool of candidates that is as large as possible seems to be the main factor in picking technologies. Even to the point of motivating rewrites. @slightlyoff So what would you recommend for someone looking to build a simple ti work with, fun, easy to use UI? @joyographic As an alternative to React, I've been very happy playing with Svelte (and SvelteKit) @joachim @joyographic There are tons of good reactive systems out there. If you like React and JSX, try Preact or Stencil. If you are building a mostly static site, look at HTML generated at build time w/ something like 11ty. Agree that Svelte is a much more modern choice. I list some others here: https://infrequently.org/2023/02/the-market-for-lemons/#fn-alex-approved-1 @joyographic it doesn't matter. use react if you want but put the user first and send the smallest amount of traffic to the browser that you can @slightlyoff @pablo_fradua it's functionally no different to inline styles. it's a solution to a problem that comes from not understanding how to work with css @joyographic @slightlyoff @digifox I don’t want to speak for @slightlyoff, but i think the point is that it ISN’T more efficient, people just cite that as the reason for choosing it @slightlyoff 100%. Unfortunately I don't get a choice but to use it at work because it's the "standard".  @slightlyoff @dylanbeattie agreed. Out of curiosity, what’s your favorite? I say this as someone who more or less rage quit front end dev 10 years ago cuz it was a dumpster fire and it’s hardly improved. I see occasional interesting sparks like Elm that try to solve only a single part of the problem. @ianthetechie @slightlyoff @dylanbeattie Loving the thread! As with any frontend tech conversation, I should mention https://htmx.org as a viable alternative to the complexity that React brings. |
Gregory's Server
@slightlyoff When you say "efficiency", what are you referring to? Development efficiency or the efficiency of the code running on the browser?
I'm curious, as I can agree with one but not the other.