You can't just use HTML because browsers have mutated to fat clients for a X like protocol which is a mishmash of html, css, js, and whatever else someone's decided to throw on top of it.
We're using screwdrivers as axes and everyone is acting like this is some type of acceptable outcome.
You can't make as rich an interavtive experience as you can with JS, but you never could.
These are some of the JS functions HN runs in the background, on top of the CSS:
function $ (id) { return document.getElementById(id); }
function byClass (el, cl) { return el ? el.getElementsByClassName(cl) : [] }
function byTag (el, tg) { return el ? el.getElementsByTagName(tg) : [] }
function allof (cl) { return byClass(document, cl) }
...I think the parent poster was referring to how much noise there is, and I'd agree. The moment you start learning, just the fanciest stuff is shoved down your throat. So you just assume its the best and roll with it.
I find myself tinkering more with PHP now that I've been using js for the past 7 years.
The Oxide route is one of the better approaches for backend. "Apple" of enterprise OSS, but I think their ambitions are too small. Prefab containers full of seamless and modular amounts of each food groups: CPU, GPU, RAM, SSD, HDD, interconnect, and uplink all in and managed. Not rack-up but dirt-up and totally managed offering IAM, VMs, 12factor PAAS, serverless, volumes, and object storage with multitenancy, accounting, security, data lifecycle, config management, appropriate redundancy, and other cross-cutting concerns harmonized in a way that is necessarily managed but sufficiently customizable.
Frontend, the trick is standardizing on the least fragile tools that are widely used enough. Churn on tools and dependencies is a distraction and a time waster.
This is the nuts of it.
There is no need to follow the trend du jour, it is some fallacy that you're describing that somehow it's easier then than today when the technology is largely backwards compatible.
There’s lots of reasons a Google is impossible to start today (the main one being “Google exists, whatever the next explosive startup-to-giant is [0], it will look nothing like Google, and such things aren’t cookie-cutter, each is sui generis), but the explanation you offer above is… unconvincing as a bare conclusion, but maybe could be fleshed out with more description and support.
> Not sure what the solution is but we need fewer sheep in development and less permissive licenses so developers doing unglamorous work can capture more of the value.
The two halves of this sentence are in tension, and the first seems more reasonable than the second.
> Look at how much effort it took to write the cgi-bin scripts google started with vs whatever flavour of the week JS framework you have to use now.
You don’t have to use a flavor-of-the-week JS framework in place of cgi-bin scripts. (And, in some ways, the lowest-friction backend options are lower friction to get up and running than cgi-bin scripts on a server you set up, because you’ve got things like “serverless” FaaS hosts.)
[0] OpenAI?
Looking at the state of open source software today a google is simply impossible because the ecosystem has rotted from the inside.
Look at how much effort it took to write the cgi-bin scripts google started with vs whatever flavour of the week JS framework you have to use now.
Not sure what the solution is but we need fewer sheep in development and less permissive licenses so developers doing unglamorous work can capture more of the value. There's a reason why every shop which supports massive open source projects is running away from legacy licenses as fast as they can and that reason is Amazon.
If you don't care about developers from the user side of things it's just as bad. The GPL in the age of cloud services does as much to protect user freedom as the MIT license did in the 1990s.