5 ms·

Yep, I tried front-end way back and hated it. Can you mention some other tech that should be covered in my re-evaluation?

teryyy 6y ago
TailwindCSS, Svelte, and Snowpack are probably worth a shot, all pretty much bleeding edge
- 1_player 6y ago
  The problem with front end development is that it's all bleeding edge and nothing that's stable and is improved for more than one hype cycle. Suggesting more bleeding edge tech is not a good argument to demonstrate that front end development got better. In my experience, it changed, it improved on many aspects, but it's still a pile of unstable crap running on the power of hype and over-enthusiastic junior engineers.
  - qbasic_forever 6y ago
    I get your frustration (and shared it a year ago), but you're really doing something like react a disservice by dismissing it like that as just hype. On the contrary react has evolved significantly over the years, introducing things like JSX (templating HTML + JS in the same file, now standard in every other component system), build tooling like create-react-app (zero config tooling, now standard in every JS build tool), functional component model with hooks (now copied by every other component library), and on the horizon stuff like server-side components. It has only gotten more stable, more proven, and easier to use over time.
    - 1_player 6y ago
      I am not dismissing it. I appreciate what React is doing, and it is indeed revolutionary. My point is that there has been a lot of evolution and improvement, but developer sanity is one of the few things that have just gone worse. It is dishonest to look only at the bright side of modern frontend coding and forgetting the node_modules nightmare, the incredibly obtuse language, the slow compilation times, the myriad of CSS frameworks that are invented because CSS is an absolute pain in the arse for modern design, etc., the fact that every JS tool either is core to the whole circus and is maintained by a stressed out, underpaid person in the middle of nowhere, or has a minimum of 5k Github issues and a major version released every 3 months, and every time you get a problem you're told "haven't you tried the new alpha release of the big rewrite we've just started?". It is absolute madness, not just frustrating. But yeah, React is cool.
protoduction 6y ago
I'm really biased as I am building it myself: Starboard Notebook puts literate programming into the browser without a lot of magic. It's amazing what browsers can do these days without any build step! [0]: https://starboard.gg https://starboard.gg
qbasic_forever 6y ago
React & next.js are worth evaluating. It's a super powerful full stack with everything from server side rendered data (REST APIs even) all the way to complete client-side SPAs, all using the same react components. It's a slick little swiss-army knife that can make anything from a basic blog with markdown content, to a static marketing page, to a full client-first PWA experience. There is sadly almost nothing else really like it right now. MDX is another neat step, it's markdown + react components and fits in very well with next.js: https://github.com/hashicorp/next-mdx-remote https://github.com/hashicorp/next-mdx-remote Vue, and Svelte to just get an idea of what other components systems are like. They're all equally capable and just have different tradeoffs and styles. Keep an eye on Svelte in particular as its next.js-like system (SvelteKit) is working on a major revamp to be serverless-first and is quite interesting. Once you learn one component system it's easy to switch between them all--they're all cribbing and building on top of each other's ideas. The whole space is innovating in a great way. Web components are good to learn and compare to component frameworks above. It's still a changing space but points to a nice future where we can all just publish and share components.
- pandemic_region 6y ago
  Yet none of the stuff you mention here will live beyond the next two years. Fun is as much as i would describe the js ecosystem, for more boring corporate use banking on these is the equivalent of getting a trojan horse in the application portfolio.
  - johnfn 6y ago
    React won't live past two years? What? I'd be willing to stake money on React, Vue, next.js and Svelte surviving at least 2 more years - roughly in descending order of probability, though all three well above 50%.
  - asattarmd 6y ago
    React was released in 2013, that’s 8 years.
    - 6y ago

teryyy 6y ago

TailwindCSS, Svelte, and Snowpack are probably worth a shot, all pretty much bleeding edge

1_player 6y ago
The problem with front end development is that it's all bleeding edge and nothing that's stable and is improved for more than one hype cycle. Suggesting more bleeding edge tech is not a good argument to demonstrate that front end development got better. In my experience, it changed, it improved on many aspects, but it's still a pile of unstable crap running on the power of hype and over-enthusiastic junior engineers.
- qbasic_forever 6y ago
  I get your frustration (and shared it a year ago), but you're really doing something like react a disservice by dismissing it like that as just hype. On the contrary react has evolved significantly over the years, introducing things like JSX (templating HTML + JS in the same file, now standard in every other component system), build tooling like create-react-app (zero config tooling, now standard in every JS build tool), functional component model with hooks (now copied by every other component library), and on the horizon stuff like server-side components. It has only gotten more stable, more proven, and easier to use over time.
  - 1_player 6y ago
    I am not dismissing it. I appreciate what React is doing, and it is indeed revolutionary. My point is that there has been a lot of evolution and improvement, but developer sanity is one of the few things that have just gone worse. It is dishonest to look only at the bright side of modern frontend coding and forgetting the node_modules nightmare, the incredibly obtuse language, the slow compilation times, the myriad of CSS frameworks that are invented because CSS is an absolute pain in the arse for modern design, etc., the fact that every JS tool either is core to the whole circus and is maintained by a stressed out, underpaid person in the middle of nowhere, or has a minimum of 5k Github issues and a major version released every 3 months, and every time you get a problem you're told "haven't you tried the new alpha release of the big rewrite we've just started?". It is absolute madness, not just frustrating. But yeah, React is cool.

protoduction 6y ago

I'm really biased as I am building it myself: Starboard Notebook puts literate programming into the browser without a lot of magic. It's amazing what browsers can do these days without any build step! [0]: https://starboard.gg https://starboard.gg

qbasic_forever 6y ago

React & next.js are worth evaluating. It's a super powerful full stack with everything from server side rendered data (REST APIs even) all the way to complete client-side SPAs, all using the same react components. It's a slick little swiss-army knife that can make anything from a basic blog with markdown content, to a static marketing page, to a full client-first PWA experience. There is sadly almost nothing else really like it right now. MDX is another neat step, it's markdown + react components and fits in very well with next.js: https://github.com/hashicorp/next-mdx-remote https://github.com/hashicorp/next-mdx-remote Vue, and Svelte to just get an idea of what other components systems are like. They're all equally capable and just have different tradeoffs and styles. Keep an eye on Svelte in particular as its next.js-like system (SvelteKit) is working on a major revamp to be serverless-first and is quite interesting. Once you learn one component system it's easy to switch between them all--they're all cribbing and building on top of each other's ideas. The whole space is innovating in a great way. Web components are good to learn and compare to component frameworks above. It's still a changing space but points to a nice future where we can all just publish and share components.

pandemic_region 6y ago
Yet none of the stuff you mention here will live beyond the next two years. Fun is as much as i would describe the js ecosystem, for more boring corporate use banking on these is the equivalent of getting a trojan horse in the application portfolio.
- johnfn 6y ago
  React won't live past two years? What? I'd be willing to stake money on React, Vue, next.js and Svelte surviving at least 2 more years - roughly in descending order of probability, though all three well above 50%.
- asattarmd 6y ago
  React was released in 2013, that’s 8 years.
  - 6y ago