Alex RussellIf you want to understand just how badly the JS-industrial-complex has messed up the web, try to view-source the NYT.
The "this site is faster in the app" prompts have the benefit of truth *because they broke the website with React*
"but Alex, it's not fair to blame React..."
Ok!
Then consider this the world's highest profile example of why Next.js is a danger to your team and your business. If the grey lady can't figure out how to lazy load images using Next with less than 1.2MB of JS, why do you think your midsize e-commerce site will fare better?
That bright young boot camp grad telling you to rewrite your functional WordPress system in "RSC"? Straight-up liability.
Go make them do a bake off of a simple blog, or an analysis of your users by device class. Do not let them code until they get regrounded.
nrkYes, that's right folks. With the power of Next.js, a full time staff of JS experts, white-glove service from your technology supply chain, and the benefit of wealthier-than-average users with faster phones, you too could achieve..."needs improvement":
https://treo.sh/sitespeed/www.nytimes.com?metrics=lcp%2Cr&formFactor=phone
But I'm sure NYT is an outlier.
Just gonna quickly check in on the BRANDS BRANDS BRANDS from the Next.js showcase site...
https://treo.sh/sitespeed/www.lg.com?metrics=lcp%2Cr&formFactor=phone
https://treo.sh/sitespeed/www.perplexity.ai?metrics=lcp%2Cr&formFactor=phone
https://treo.sh/sitespeed/solana.com?metrics=lcp%2Cr&formFactor=phone
https://treo.sh/sitespeed/openai.com?metrics=lcp%2Cr&formFactor=phone
https://treo.sh/sitespeed/www.wayfair.com?metrics=lcp%2Cr&formFactor=phone
https://treo.sh/sitespeed/www.notion.so?metrics=lcp%2Cr&formFactor=phone
https://treo.sh/sitespeed/www.target.com?metrics=lcp%2Cr&formFactor=phone
Fabrice Desré@slightlyoff I would be nice to also have a gallery of fame, not just of shame. Need to show off how good the web can be!
@fabrice This is pretty good:
@slightlyoff notion has 6.5s LCP to show a freaking static landing page that barely have any significant animations ? how even ?
@nrk9819 Behold the power of a fully operational dea...er...metaframework!
Avalynn Gunn 🇵🇸@slightlyoff Working on performance for a Next.js and Styled Components stack feels like bailing water on the Titanic.
But after 2 years we're finally moving off SC.
@clarkgunn The whole "CSS-in-JS" delusion needs to be taught to future generations of web developers as an object lesson in devastatingly incurious hubris.
The proponents of these things *were warned directly by browser engineers* that they would scale like leaden balloons, and yet they persisted, downplaying the predictable consequences, never apologizing.
Much in the self-sanctioning style of Brexit, they nerfed perf and never won the purported gains.
A CSSexit, if you will.
@slightlyoff I'm hoping to do a write up after migrating this whole component library off Styled Components.
@clarkgunn Would read/retoot.
Josh Tumath@slightlyoff @clarkgunn I would also read!!
@slightlyoff I also place the blame for CSS-in-JS on React. React wanted to own the DOM and everything but didn't want to handle styles.
Vue and Angular both shipped style solutions (which were technically CSS in JS, but shipped as actual CSS) so this was a non-issue in those ecosystems.
CSS-in-JS stepped in to fill the void, to catastrophic results.
The author of SC is a former employee of my company and I feel like I've been cleaning up their mess since I joined.
Aleksandar Vacić@slightlyoff heh…me in 2017, remembering another life as web developer:
“…I witnessed how simple, declarative world of HTML/CSS and original baby Javascript morphed over the years into increasingly complicated quagmire of “advanced something” until most front-end developers forgot how CSS selector specificity work. How property value inheritance works. Developers couldn’t be bothered to learn it because apparently they can’t do shit without variables.”
Jon@slightlyoff have you ever looked at washingtonpost.com? I had a look at their window.__NEXT_DATA__ object and then just laughed. I think I saw employees internal network IDs in it.
pavo
Tane Piper ⁂@pavo @jon_nunan @slightlyoff I was talking to a few peers at DevOpsDays Amsterdam yesterday about how we have an entire generation of developers who don't know the fundamentals - they learned `npx create-react-app` is how you build every application - and now they are at some of these prime websites building core system.
I got a few nods of agreement.
ArtemesiaOn the plus side, the slow load&render speed of the NYT makes it easier to manually defeat their paywall. Bloat on, bothsidesering editorial board!
@artemesia Would it surprise you to find out that they paywall is entirely client-side JS, and that if you block, like, 2 script files, you can read the whole site w/o a login?
Discovered this while trying to make an extension that just lightened the dang thing up.
seismo!allegra!utzoo@slightlyoff @artemesia are those two script files on a specific domain that could be blocked on a pihole?
Dang, wonder if that's recent. I've noticed that they've been iterating on their paywall.
Dusty Pomerleau@slightlyoff In terms of priorities, what NYT cares about is coming up first in search. The leaked Google documents give a lot of insight into how domains get blessed for "siteAuthority" despite failing on other metrics. It's sad, because young sites have been led to believe they can climb the rankings by obsessively tuning their lighthouse score. NYT is the flagship example showing that's not the case.
@dpom A lot of the stuff I've read about the leak is overheated, bordering on conspiracy theories.
Separated from that, it's disappointing how little weight web perf and overall usability have been given by search engines over the years. AMP was the biggest intervention, and it was terribly pitched and executed, breaking the idea of "one web" and randomizing URLs.
Search engines and browsers need to do much more.
@slightlyoff Mainly what we learned from the leak is that Google hasn't been truthful in the media about how their product works. Well-known domains do get preferential treatment, there is some sandboxing of new sites... These things have long been suspected, so I think people are more upset about misleading statements than they are about how search really works.
@dpom I'm more upset about how web search quality is continually degraded by AI and SEOs 🤷♂️