I read this article and thought YEAH!! Then noticed once I got to the comments that it's from 2016. Which is really depressing because all of the problems mentioned by the author have gotten worse, not better in the last few years. Ironically, the new frontend Twitter is currently beta testing is built using React Native > Web tooling which is laughably non-semantic in the extreme.
I don't know what the ultimate answer is to this, expect that perhaps standards-based Web Components can get good enough that devs will rely more on those approaches than ginormous JS frameworks that take over everything. In the meantime, the web is not a friendly place for old-school devs lacking a JS-ALL-THE-THINGS! mindset (like me!).
Trying out a job site I hadn't used before, I found a position at a local startup. A run of the mill startup, but well funded and with a non-amateurish brand feel.
Despite being a few simple pages, with not much interactive features and no experimental layout stuff, their website required JS to load — a 3.7MB bundle.
I briefly entertained my pedant side, wanting to send them an email, but I didn't. Let the devs play, let the execs run their racket, let the market sort out whatever.
-
Thinking about the startup's product, it sounds like a convoluted implementation of an existing product from a multi billion-dollar company. So maybe it's turtles, bullshit jobs and inscrutable schemes all the way down.
"The reply button, for example, focuses and expands the textbox below. You can’t do that without some scripting"
-- Can be done with just CSS :target
"Similarly, the button does an action behind the scenes, which is iffy since you could replicate it with a full page load"
-- Should just be a normal link with a full page reload if JS disabled. Perhaps this could also be a new use of the <a ping> attribute (without a href).
Expanding the textbox can, but how would you set focus to another element with css? Genuinely curious, I have just learned about :target and am having this exact scenario in a project of mine.
I don't know what the ultimate answer is to this, expect that perhaps standards-based Web Components can get good enough that devs will rely more on those approaches than ginormous JS frameworks that take over everything. In the meantime, the web is not a friendly place for old-school devs lacking a JS-ALL-THE-THINGS! mindset (like me!).