Thread 105946460

The under 14kB edition.

>Free beginner resources to get started with HTML, CSS and JS
https://developer.mozilla.org/en-US/docs/Learn - MDN is your friend for web dev fundamentals
https://web.dev/learn/ - Guides by Google, you can also learn concepts like Accessibility, Responsive Design etc
https://eloquentjavascript.net/Eloquent_JavaScript.pdf - A modern introduction to JavaScript
https://javascript.info/ - Quite a good JS tutorial
https://flukeout.github.io/ - Learn CSS selectors in no time
https://flexboxfroggy.com/ and https://cssgridgarden.com/ - Learn flex and grid in CSS

>Resources for backend languages
https://nodejs.org/en/learn/getting-started/introduction-to-nodejs - An intro to Node.js
https://www.phptutorial.net - A PHP tutorial
https://dev.java/learn/ - A Java tutorial
https://rentry.org/htbby - Links for Python and Go
https://quii.gitbook.io/learn-go-with-tests - Learn Go with Tests

>Resources for miscellaneous areas
https://github.com/bradtraversy/design-resources-for-developers - List of design resources
https://www.digitalocean.com/community/tutorials - Usually the best guides for everything server related

>Need help? Create an example and post the link
https://jsfiddle.net - if you need help with HTML/CSS/JS
https://3v4l.org - if you need help with PHP/HackLang
https://codesandbox.io - if you need help with React/Angular/Vue

/wdg/ may or may not welcome app development discussion. You can post and see what the response is.
Some app technologies of course have overlap with web dev, like React Native, Electron, and Flutter.

We have our own website: https://wdg-one.github.io

Submit your project progress updates using this format in your posts, the scraper will pick it up:

:: my-project-title ::
dev:: anon
tools:: PHP, MySQL, etc.
link:: https://my.website.com
repo:: https://github.com/user/repo
progress:: Lorem ipsum dolor sit amet

Previous: >>105872088

>>105946460 (OP)
>Why your website should be under 14kB in size
>Because of the TCP slow start algorithm
Doesn't this apply only to HTTP/1.1? HTTP/2 and HTTP/3 should have solved this.

>>105946520
>Similarly to HTTP/2, there is a notion that HTTP/3 and QUIC will do away with the 14kB rule — this is not true. QUIC recommends the same 14kB rule.
https://endtimes.dev/why-your-website-should-be-under-14kb-in-size/

>>105946569
How can QUIC authors recommends the same 14 kB rule? The protocol doesn't even use TCP, it can't be affected by the slow start algorithm, QUIC is built on top of UDP.

>>105946460 (OP)
lol
lmao even

>>105946622
>How can QUIC authors recommends the same 14 kB rule?
Read the docs.
https://datatracker.ietf.org/doc/id/draft-ietf-quic-recovery-26.html#section-b.1-2.2

>www.www3.planetcom.co.uk
>h30434.www3.hp.com
Why do you guys do this?

>>105946460 (OP)
Best I can do is 30MB

any good resources for learning the fundamentals of page layout/composition? my stuff always looks slightly off. feel like it could use the space better.

>>105946768
elements of typographic style

>>105946768
Copy popular web apps that you like and that do something similar to what your web app should be doing. They probably know what they're doing. And just build shit. You learn the best by doing the work.

>>105947169
ah yes, we need more of the internet to look the same.

>>105947188
Fact is that people like web apps that are familiar because it makes them more intuitive to use.

What's the least cancerous registrar for .moe domains that won't fuck me over in one way or another? All else being equal I'd prefer a European one, but most importantly I want a reliable one.

>>105946460 (OP)
Does anyone know a good way to unbind event listeners that use anonymous functions that curry other functions? I've got a little spa that hosts a couple of games and It leaks around 15mb of memory every time you load a game without refreshing the page and I'm assuming it's because the call backs are keeping heap allocations alive when they should be dropped.

>>105947169
solid idea. wish there was some sort of reference book for it though.
>>105946792
this might be a good place to start. thanks anons.

Are there any FastAPI saaars itt

>>105946682
I've read it. It references RFC6928 which is for increasing TCP segments from 2 to 10.
The 10 burst TCP segments was done as a mean to not overwhelm internet infrastructure and to not waste packets. The RFC6928 on page 5 also mentioned that web browsers like to go around the TCP congestion mechanism by opening multiple TCP connections to speed up sending and receiving documents, this is kind of similar to multiplexing but not really the same.
This RFC does not mention UDP at all, it's for TCP only.
The multiplexing of QUIC and HTTP/3 is their core functionality, there is no congestion control, packets are sent as soon as they are requested with no limits to a first batch of initial packets.
The 10 burst requests apply only to TCP, thus to HTTP/1.1 and HTTP/2, it does not apply to HTTP/3 as it uses UDP/QUIC.
I don't know what the QUIC authors were smoking when they recommended the 14 kB rule.

tailwind is based

>>105946460 (OP)
I'm a total noobtard in webdev and want to make a simple SPA that takes a log file, stores its metadata in a browser-local db, then shows a table that allows the user to filter it by cols, and allows exporting to a file.
How should I start? React? Preact? Solid? Clojurescript?
I want it to work five years in the future (because browsers don't change!) and I don't want to wrestle with retarded doo-hickey special-isms of whatever faggy framework I'm forced to use. It better be straightforward and fast.

>>105947801
I'm actually saying this unironically, just vibe code it, it's good at this sort of basic stuff.

But in general you really don't need a framework for this. This sounds like something that's easy enough to do with vanilla JS. Maybe jQuery for some convenience functions. Just make the UI in HTML (tailwind and/or bootstrap may help) and add onClick handlers.

>>105946768
<table></table>

>>105947859
I want to write this just so I know what I'm doing, because I'm a NOOBTARD.
And thanks, might just use JQuery. And no tailwind, class vomit cancer, doesn't run well on slow computers.

>>105947880
how many people do you think write html these days?

>>105946737
And my son, FAT32

>>105947859
What database can I use here? WASM SQLite is too heavy.

>>105947801
>>105947882
>How should I start? React? Preact? Solid?
No. Vanilla JS. A lot less mess and boilerplate, then you can design it later (with or without frameworks). What you're talking about isn't really related to frontend anyway, rather just a JS application
(Parsing the log, using IndexedDB)

>>105946460 (OP)
I WILL use React and my site WILL be larger than 14 kb and I WILL NOT care.

>>105948110
Is there too much data to just store it in memory?

>>105948128
Neat, thanks. Yes, it's not explicitly "frontend" (and there's no backend either).
>>105948155
It could be like ~50M of plaintext. Is that "too much"? It could be too much to process with a browser-native solution like IndexedDB.
Assume I need to do a "select * from mytable where x = y" or something, over ~10k records.

>>105946569
>site loads instantly
Damn, I'd forgotten they could do that.

>>105948449
>Embeds duplicate stylesheet in every page, for performance theater
Fucking webshitters man. Will they ever get over their obsession with tinkertrannying?

>>105946569
>you should make web pages that don't serve any real function because it's slower for some literally whos out in the middle of nowhere using obsolete geostat satellite internet
no thanks

>>105946460 (OP)
My pages are multiple MB and they load fine on all my devices.
They're probably slow if you have a shitty device/connection, but my site is commercial, not a public service or charity. As cold as it sounds, it doesn't make much business sense to spend time and effort optimising for poor people with old gear.

>>105948967
>Turning a profit validates my mediocrity and inefficiency

>>105948884
>t. makes websites that still don't serve any function, but are 10M instead of 14K

>>105948967
fuckin hell when will you faggots get your shit in gear

>>105948884
Good post

>>105949021
Sounds based to me

hello devs, got a question, I am starting to learn React and so far I did my landing page with it, my question now is (I want to make some sort of an app) is it a good way to have the landing page in /src/landingpage and the app itself under the same src folder like /src/app ? Is this okay?

>>105949894
like this
project-root/
── frontend/
── src/
── app/
── landingpage/

rewriting my shitty go web app in spring boot, and questioning why Go even exists

>>105950144
pike was pretty clear, he wanted to make a jeet-proof language

>>105950790
When did the stereotype that Java is for jeet even start? If everyone uses the language, it's normal jeets would try to learn it but not like it matters.
I love having shit just work as I expect it, without dealing with the shit that Go made me go through
>SSE? There's this package, that doesn't actually implement the standard for some retarded reason, but here's your alternatives and each one has some stupid flaw that's a deal breaker
Why is it like this? Meanwhile, it took me a day to rewrite everything, mostly just translating logic language to language and it just fucking works how I want

>>105949894
>>105949921
If your app is going to have multiple pages and you're using a router to do that (React Router, Tanstack Router, whatever) then maybe the landing page can just be treated like any other page?
But if you prefer to keep it separate then sure, I don't see any issue with your structure.

As a third option... at my work we just use something like Squarespace or Wix for the landing/marketing/contact pages, and then the actual app lives on a subdomain.
So
www.example.com => squarespace marketing pages
app.example.com => actual webapp

>>105950832
sometime shortly after the signing of the immigration act of 1990 that opened the floodgates on the cesspool

>>105946460 (OP)
>why your website should be under 14 kb
but sar... any production codebase has 10,000s of modules

>>105950832
Way to out yourself Pranjeet.

>>105950832
What do you even mean? Go is simpler than Java

>>105950838
this is the way. marketing should be separate from the application. The marketing page should be attached to some sort of CMS so you can edit content without having to rebuild. It should also be server side rendered for SEO.

Then the webapp can be whatever you want. Personally I opt for a SPA with react and a Go backend. Either embed react to be served from Go itself or throw it up on Cloudflare as a static asset for unlimited free hosting. I do the latter to decrease outbound bandwidth significantly.

Just have *something* load and display with that 14kb.
Like Windows XP putting up the desktop but not responding to anything immediately.

>>105947969
Neocities makes html and tables cool again

Has anyone vibe coded a bare minimum product and then rented it out to a business? I have an opportunity that I can probably leverage due to the cheapness of the people running the operation and unwillingness and incompetence of the people who work there from coming up with a solution

I have never actually coded anything, other than HTML sites, tinkering with setting up servers, self hosting, stuff like that.

Or is vibe coding literally just a marketing meme made up to drive AI subscriptions? I am led to believe that you can literally sit down on a laptop, type in what you want to do and then make it lead you to the outcome of a working product. Sounds too good to be true to me.

>>105946460 (OP)
c programmers can WHAT
https://nigeltao.github.io/blog/2020/jsonptr.html

>>105948154
It shouldnt matter after initial load

>>105950992
Durgasoft seems based. I would trust them with my enterprise solutions architect needs. I'm rating them 5 bowls of vindaloo and lets throw in a couple of garlic naans just because I really like Mr. Nagoor's stage presence

>>105946460 (OP)
>14kb
thank you, very cool

Finally learned MERN I think after many hours of floundering in GPT but now it's late and I have to start drinking if I wanna kill the vyvanse high and sleep
Oh, woe...

>>105940562
don't wanna jinx it but it went well

I'm supposed to get a response next week

>>105950144
>>105950790
>>105950832
I prefer Go to Java. Java is slow and heavy and heats up my CPU - same with Java tooling like the language server. Go doesn't do these things. And Go builds much quicker too.
>SSE? There's this package, that doesn't actually implement the standard for some retarded reason, but here's your alternatives and each one has some stupid flaw that's a deal breaker
I think the Go ecosystem probably isn't as mature as some other ecosystems. But usually I can find packages I need. I've not tried server-sent events though.

Page 8

>>105948449
For real though why did it load so fast? I'm not a web dev. I guess because it only requests one other resource (some analytics js) and has that fetch deferred?

>>105954993
it literally describes in detail why on that page, fuck

>>105950838
>>105952624
This sounds better, I think I will make it like you guys recommended, have the landing page its own React project (will use Strapi with it) and the app will be a React project itself too.
project-root/
-- landingpage
-- strapi
-- app

>>105950838
>>105952624
This sounds better, I think I will make it like you guys recommended, have the landing page its own React project (will use Strapi with it) and the app will be a React project itself too.
project-root/
-- landingpage
-- strapi
-- app
-- frontend
-- backend

>>105946460 (OP)
>mine is 17kB
It's joever

>>105954993
Because it has no images, it uses very minimal CSS and it doesn't use JS for anything other than some analytics.

Any of you work with web components on a regular basis? I got a couple of offers to work with Lit and I'm tired of React.

>>105955949
>web components
I'd rather work at the local capybara park

jokes aside, it should be alright. web components aren't that bad.

>>105946460 (OP)
delusional. limiting yourself to 14kB would be a stretch even 25 years ago.

>>105950790
Reflection-based code is hella ugly compared to normal code
Show the equivalent in Java, if it even has that
t. contributed to Hugo a tiny bit

>>105946460 (OP)
14 KB is something I can work with but I like shipping a font or two and I can’t always pick a variable font

Do you guys like Neocities?

>>105956823
idk what that is, but I remember my first geocities pages, like 20 years ago if not more, that was hella cool, with the robot walking gifs and marquee, I think CSS didn't even exist yet, or at the very least I wasn't aware of it, it was all html styling

>>105956798
i think 14k for everything limits what kind of pages you can make quite a bit, but the critical take-away is to at least try to make the page /useful/ at 14k, like ensure the first 14k shows the user the text they went there for, along with basic formatting. custom fonts can come later

>>105946460 (OP)
I've read this article before but did not understand if it applies to the seperate parrallel requests or individual ones.

Like should my HTML + CSS be under 14KB, or should my HTML be under 14KB and my CSS be under 14KB, but together they can be more?

>>105957210
you should probably learn a bit about how networks work before getting into a profession that focuses on network services...

>>105946460 (OP)
>your website should be under 14kb!
>no, not mine though, *YOU* deal with that!

>>105957322
Nvm, looked it up, TCP slow start is done on every connection, even parrallel ones done against the same server. So the performance is as follows:

HTML + CSS (Inlined styles) > 14KB:
Base Latency + SS

HTML + CSS < 14KB:
Base Latency

HTML < 14KB
AND
CSS < 14KB:
Base Latency / 2 (Parrallel)

HTML > 14KB
AND
CSS > 14KB:
(Base Latency + SS) / 2

HTML < 14KB
AND
CSS > 14KB:
(Base Latency + SS) / 2
(Assuming CSS is blocking, it will bottleneck even if the HTML is under 14KB)


I know how networks work, I was asking cause I did not feel like looking it up and wanted a quick answer from an Anon who has thought about it before, seems not many care about such theoretical micro optimization questions. And no, I am not looking for a job in web dev thank you very much, this is /wdg/, you can do your recruiter roasties should go to /twg/ or /utwg/.

>>105957489
Compressed it's 10KB, your browser does support compression and sends the headers for it right Anon?

>>105957633
What if I use browser that doesn't support compression?

>>105957640
If you actually are, please tell me which it is, sounds interesting, caniuse says gzip compression "is supported in effectively all browsers".

I'd guess surf or lynx or one of the other minimal browsers, but the network debug pic you posted looks like chrome/firefox.

Been working my website (https://csbook.club/) more.

I found a dump off all the /prog/ textboard posts from 2008-2014. I've been working on porting them over just because I used to spend quite a bit of time there back in the day.

I tried an experiment with migrating a thousand threads and this webm was the result. In total the dump has 1018696 replies to 31362 threads.

It's been pretty fun and a lot easier than expected.

>>105950832
skill issue

>>105957633
so you're saying if i have a retarded stack like HTMX i can optimise for this?

>page load it's just template <14kb
>HTMX.js <14kb (i hope)
>HTMX makes a bunch of <14kb calls to each section because i made it so
>CSS <14kb
i get base latency on all but at the massive cost of everything popping in

and you save what? milliseconds?

about what some of you're saying, what's the benefit of loading an UI fast if it isn't even interactive yet, the user will get frustrated for sure
I'd get it if it was an obvious placeholder like (this is how it looks but also letting you know that it is currently loading and not interactive)
that's better UX, no?

>>105946460 (OP)
deny port 80 accept port 70

>>105957979
Correction for my previous post:
HTTP/1 Makes multiple connections which each have a slow start
HTTP/2 or HTTP/3 Make one connection and multipplex it to serve multiple simultaneous request, requiring only 1 initial slow start.

In our case it does not matter though, cause your blocking resources are all gonna be requested at the same time, that is to say at the very start of page load. So you are still getting cucked if atleast one of your blocking resources is above 14KB.

As for your HTML/HTMX/CSS question, HTMX is more then 14KB gzipped, but only slightly, if you treeshook the parts you did not need it would be under 14KB gzipped, but I don't think HTMX is blocking to begin with, maybe the more advanced features of HTMX? But for the basic stuff which I know I don't see why it would be blocking and critical to get to the user in initial page load.

You mention HTMX is making calls to sections which are each <14KB. So it's requesting parts of your webpage from the backend, guessing some kind of templated stuff with dynamic values? If so, you could maybe lazy load them if they are not on the initial view, or use a placeholder for them, or rethink if you really need sections of your page to be dynamically fetched from backend with HTMX at the very beginning of page load which sounds kinda odd.

You definitely don't want critical things popping in, but you also want to save those milliseconds if you can without making important things pop in.

>>105957999
IRL you first see something and then you interact with it, the minimum time between you seeing something for the first time and interacting with it is far greater then the time it will take for your page to load interactivity after it has loaded the content. Your eyes are much faster then your hands and mouse, so from a UX perspective it is absolutely better to load the UI fast at the cost of loading the interactivity with some minor delay, then it is to have the UI load slower but in sync with the interactivity.

Only exception to this is massive framework sites which take so long to load that when they do load the user is so exhausted they will get mad if there is no interactivity right away on top of that, but that rarely happens, most placeholders are for data anyways, not interactivity.

>>105957999
better yet, super fast interactive (so load js first, no images at this stage, animated gradients, text and hyperlinks) mockup, with minimal required content for interaction while the page gradually loads the rest

>>105957979
>"and you save what? milliseconds?"
The point of the website posted was that depending on the user's connection, they could be saving up to a second in the worst of cases, which is something they absolutely will notice on initial page load.

>>105958079
doesn't seem optimal, but it has obvious workaround, load minimal interactive version (js,first) + html + css, then offload the rest for later

>>105958093
>>105958105
I feel like you are trolling because you are describing exactly how big JS frameworks tend to operate with placeholders, but the thing there is that they are using JS to create the entire UI in the first place, so JS becomes the bottleneck for interactivity and layout. In any case, the user has to see what they want to interact with before they can interact with it, so UI is more important then interactivity.

Who is going to click on your website and try to be able to click on a JS button in under 100ms? Their eyes might detect 100ms latency, and when you are playing a realtime game you will detect a 100ms delay in clicks, but when you first load a page, even one you know by heart, you will not get cucked by a 100ms delay in interactivity, just moving your mouse to the button will already be enough time to load the JS.

All the placeholder stuff is for dynamic data fetching from the backend and then sorting/filtering it on the client side, which does take alot of time for certain data. But just regular interactivity does not need a placeholder.

>>105958165
>In any case, the user has to see what they want to interact with before they can interact with it
why? why not load MINIMAL INTERACTIVE VERSION, then let the rest lag behind?
if your minimal version can work without js, so be it, if it needs minimal js version, so be it
>under 100ms
realistically nobody
but if we're talking 1s+ then yeah, I can see the user thinking (what kind of retarded n*g*er shit is this?)

>>105957187
Pretty much the same, but watered down and filled with you know who
https://glitchedguts.neocities.org/

>>105958684
pretty basé, can yoou still js/css hack it just to hide the yahoo ad shit?

page 8, really n*g*as¿

Just code the website in ASM, problem solved

There is something really nice about fast websites and web apps. They are much more fun to use.

>>105960306
google lighthouse has psyopsed manager types that fast = moar money from users
I for one don't have the data to verify it, but it kinda makes sense, but it's find to picture that kind of user that will ctrl w your website on 100ms+ load time

>>105946460 (OP)
How would you do this using only vanilla JS, HTML, and CSS? Would you just load the bare minimum in the first request and then automatically load the rest (in bigger requests)? But won't this make caching harder? How do you 304 cache everything (as the same request) if you have done several requests to build the page?

>>105961113
good questions for sure, I too would like to know

>>105961113
You wouldn't

Websites are bloat

Write your thoughts in a journal and draw art and then publish them to a plain HTML website if you want.

But even that is bloat, astral project and write your thoughts in the ether to make them timeless

>>105957753
I've managed to get every /prog/ thread migrated over and indexed, seems to be working well. Going to update the site tomorrow with it.

>>105961388
your thoughts are bloat because they have no use to anyone

>>105958102
>>105957633
>>105957210
Lmao I just read the article and it's examples are retarded (pic rel)
>thirdworlders
>people at music festivals
>bad stuff???

Ah yes, when I'm at a music festival the first thing I'm going to do is go to anons blog and admire how it took 300ms instead of 600ms to load his blog on why op is a fag

I'm looking for a means to easily transfer a working server environment from one Windows PC to another. Unfortunately, Apache and most SQL databases are NOT portable. What's the best approach?

>>105947188
You could always just pull from captures of older websites, which do look different to new websites.
The fact more websites don't allow a custom frontend is baffling, honestly.

Everyone says Laravel is the easiest and fastest framework to learn but their documentation fucking sucks. It's just a bunch of pages of dense information about every tiny feature. Why is there no straightforward tutorial showing you how to build a website from start to finish that uses all the major features of the framework, like what React has?

>>105963268
Yeah that's the difference between documentation and guides/tutorials.

There used to be bootcamp.laravel.com which focused on the tutorial side of things, it was great and got you straight in to a step-by-step twitter clone project but apparently it's getting re-worked.

In general though, Laracasts is a great resource, this guy is the GOAT of PHP tutorials:

https://www.youtube.com/watch?v=1NjOWtQ7S2o

How do you deal with a one to many relationship in SQL and fetching data? I got posts and each post can have multiple files, so what I currently do is just fetch all posts, then fetch all files by post ID and then iterate and map them. AFAIK it's pretty efficient, I'm just curious what other people think

>>105946460 (OP)
>Why your website should be under 14kB in size
Because it's 2025 so no one will ever look at it but you so you should make it easy to look at a lot

>>105964403
>so what I currently do is just fetch all posts, then fetch all files by post ID
that's retarded, and would fail you at any SQL heavy interview. This would work if you were using NoSQL but in relational dbs you always minimize the amount of queries by using joins as much as you can

>>105961113
bumping this question from another anon
>>105964403
yes, as the other anon is saying you try to do everything with a single query, joins

>>105964920
>in relational dbs you always minimize the amount of queries by using joins as much as you can
True, except if you use SQLite.
https://www.sqlite.org/np1queryprob.html

>>105965045
it might not be as noticeable since SQLite is local not client-server, but in great majority of cases you're not gonna work with SQLite, but with PostgreSQL or some cloud variant like Aurora (unless you're a mobile dev or you make desktop apps)

>>105965120
nta, don't know the statistics, but I figure MySQL or MariaDB might be very popular, more so than PostgreSQL in most cases

>>105965142
missed my point, just saying that client-server dbs are a great majority in web-dev / backend jobs

btw aurora supports a mysql variant ;)

>>105965120
But you can make 100s or even 1000s of queries for a request if you're using SQLite (on the same server) and have set up the appropriate indexes. You literally won't notice a difference. That's the point of that article.

>>105965259
that sounds great if you have a monolithic app on a single server with lower / medium traffic, but in many cases there are many app instances / microservices, not to mention patterns like CQRS, none of which are possible with local databases

and still great majority of companies are still going to use a client-server db even if sqlite could be used instead

>>105964403
json_agg
https://medium.com/@clementgrimault/optimize-the-way-you-fetch-relationships-with-postgresql-7711fe6457d2

will making small web based games like wordle help me get hired or just show that im an unprofessional embarrassing autist ?

>>105965708
You're unemployed

>>105962065
I got the living archive of /prog/ uploaded and it seems to be working fine. I've been enjoying browsing the time capsule shitposts.

https://csbook.club/prog

>>105947783
Ever since I started using tailwind/daisyui, I understood why we can't have nice top navbars.
Using bg primary colors makes it stand out too much
Using base-100 makes it merge with the background.
The other colors are also too bright

>>105966160
??? https://tailwindcss.com/docs/colors#customizing-your-colors

>>105966187
>docs
I can't reaaad.
No, but really I wanted something that just werk OOTB to use the dozens of themes made by daisyu/tailwind

It's over jshitters bros..
https://freedium.cfd/https://medium.com/@ThreadSafeDiaries/he-rewrote-everything-in-rust-then-we-got-fired-293e3e16c2d3
>>105962208

>>105966407
yah, call me when rust works on my browser's console

>>105966407
They would save even more money than running their own servers instead of using aws.

>>105966407
They would save even more money running their own servers instead of using aws.

>>105946460 (OP)
Let's start by making webpages that are smaller than 14MB

>>105967243
kek'd

I have a dumb question.
Let's say, I have a file on a local machine the user has given me permission to load. How do I keep it loaded between browser sessions?
My use case is a local HTML front-end that replaces the HTML page on the server. There are other use cases though, like image editing.

>>105968651
>Let's say, I have a file on a local machine the user has given me permission to load. How do I keep it loaded between browser sessions?
You can store it in IndexedDB.
https://en.wikipedia.org/wiki/Indexed_Database_API

>>105968686
Oh sick! Thanks anon!

>>105969052
>>105968686
>>105968651
nta, fukken noted
>>105968686
where/how did you learn about this?

>>105946460 (OP)
>Look up WhoIs tutorials
>They always show the full info
>Use WhoIs
>Everything but registration/expiration/renewal date is hidden
I am disappointed. Is there something else I can use to find a site's owner's name?

>>105969069
>where/how did you learn about this?
not him, but you find it pretty quickly reading about webapis on mdn, and its practically twins with localstorage.

>>105969124
there's too much stuff on mdn to be reading every entry, but yeah I guess I should dig more

page 8
https://www.youtube.com/watch?v=KY6Cr6t-cuA

i should give up and just make fursuits for a living or something

>>105971058
I don't fursuits are particularly profitable. Sure, the supply is low, but good luck getting seen.

>>105967057
t. layoff by Kabir

Webdev can be replaced overnight by vibe coding. That's like the first field to get killed by AI

>>105972137
You can build any slop web app by vibe coding, but it won't be very good and certainly won't be production ready. Web dev is the most interdisciplinary CS field, so I think it will be the hardest to replace with AI.

>>105973118
I realized this when I tried to build my own web app, and now anyone claiming webdev is kill just doesn't know any better. I wager backend is more at risk

>>105972137
>Webdev can be replaced overnight by vibe codi--OHNONONNONONONONONO THEY'RE IN THE DATABASE NOW

>>105973118
It'll be production ready and you overestimate your skills. Besides, if rajeesh can get the job you bet the bar wasn't that high in the first place

>>105973372
>webdev are doubling as cyberexperts now like they're not getting owned left and right well before AI
lmao that's not even a bait, just pure idiocy

>>105973389
Literally nobody said they weren't getting owned left and right before AI. And security is absolutely a responsibility of a web dev.

>>105973407
You just posted a clown who didn't think about security at all. Any LLM would gladly spit to you the best security practices to implement for your project. So how is that even an anti-ai argument?

>>105973420
>Any LLM would gladly spit to you the best security practices to implement for your project
yeah if you ask it to. back in march the best was sonnet 3.7 which would reward hack your project to shit with security holes
a lot of these codegen platforms like loveable and replit and built stuff to protect against these issues now too

>>105958042
1965 > 70

I wonder how much extra compression you can get with Brotli if we’re explicitly saying we care about transfer size
It says the standard Brotli encoder/decoder has 120KB dictionary of common strings
This is going to be crazy useful for something like this: https://en.wikipedia.org/wiki/Brotli#Algorithm

Realistically, what could you get on a 14kb website today?

>>105974632
direct answers

>>105974632
Plain text, no bullshit

>>105974632
Firefox says https://danluu.com/why-video/ is 12.3 KB transferred and 30 KB just sitting around
So you could have a long wordy article that’s been previously minimized and has basically no CSS
This is just compressed with gzip, not Brotli

>>105957999
>>105957979
the absolute FUCKING STATE of webshitters

>>105957999
i hate websites with loading animations
more often then not the shit i wanted to see loaded many seconds before the loading animation goes away, so i'm spending most of the time just waiting for that loading overlay to fuck off

welp, claude code is down. how am I supposed to vibe now??

>>105974763
>text shows up instantly
>a second later a loading screen covers the whole page
>it continues to load the pictures and shit you don't care about
>5 seconds later you're finally permitted to read the text
thanks obama

>>105975439
personally i use a magic wand but im sure you can use a bullet vibe in a pinch

>>105946460 (OP)
>Website should be under 14kB
>Look inside
>48kB webpage
https://endtimes.dev/why-your-website-should-be-under-14kb-in-size/

>>105974632
idk, how about a Quake demake?

https://js13kgames.com/games/q1k3

>>105974706
enjoy serving 14kb site optimised for third worlders

>>105976230
it's 11k, it's about transfer size, so compressed size counts

>>105976361
The initial request should be 14kB so that the FCP will be really good, but you can of course do several requests after that. It will be noticeably faster for everyone. Really fast websites and web apps more fun to browse and that's good for business as well.

https://www.npmjs.com/package/basedom

new vanilla javascript framework just dropped
>single file components in .html
>no build required, works as es module in the browser
>fine-grained reactivity updates with automatic dependency tracking
>no typescript, no vdom

>>105976564
Looks cool. Not having a build step is nice.

>>105974706
enlighten us, einstein

>>105976564
how would redundant js ancss would be handled then? main component? Idk, something is missing

>>105976564
If it does not tackle the whole
>state as the single source of truth
thing in its current iteration when in fact the DOM *is* state, then I don't want it. We have got more than enough solutions for that. Both in browser, compiled and/or transpiled.

People have been fooled into this insanity because they somehow think that reactivity equals "markup like" syntax. I hate all these damn counter examples, too. I mean, html/js is dynamic. You can quite literally do "element.textContent++".

I am not saying that there might not be cases where this paradigm works, but generally speaking I don't want multiple instances of the same information on my screen anyways, so why are we not able to actually work with the DOM which is a nice tree structure?

C game dev here. Since I'm obviously starving and my game is a decade from release I'm thinking of setting up a web store to have something to eat and fix the three leaks in my shed. Is WordPress with a bunch of plugins like Yoast SEO and WooCommerce still the obvious way to go? I just looked at the source for a page in the top results and was greeted with a 10k character single line lazy load script, is that how you guys do it? I'm not even joking and the site runs like shit, but Google's SEO loves it, and when in Rome, etc. I doubt making a 14kb web store relying purely on HTML+CSS+a little bit of PHP or Rust or whatever for the backend side (handling orders, notifications, sending out emails, etc) will be seen as gracefully by the reptilians that control us.
No, this post really isn't a joke. But life sure is fucking hilarious. Any help is appreciated, and much love.

>>105977497
It does work with the DOM. That's kind of the whole point of it.

Can a beginner who had previously created static web apps in the past but doesn't anymore build a decent web app in 100 total hours for commercial use? Anyone know of any good web apps I can build excluding online shops or catalogues and which could be used for commercial use? Something that has multiple functions not just one button or two options, like 3 or 4 things you can do on the web app. Also I won't be using bootstrap. In the worst case scenario I want a complete web app rather than something good but incomplete but thats the worst case scenario if things go south. Also I have to write a technical report on this and record a demo of the app.

>>105978291
Of course, it has to "work with" the DOM under the hood. All frameworks/reactive libs have to, because that's the API.

What I meant was thinking about actual alternatives that do not do pseudo-html with non standard syntax extensions. I just don't want to believe that this is the best we can do.

>>105979401
It's the best that javascript is going to get

>>105978995
this sounds like a homeworkpost but its summer and also you dont sound coherent enough to be in college. why?

>>105978995
calendar todo list schedular bullshit thing to keep yourself organized yw

>>105980343
>>105978995
nta, a bit more advanced, code/code snippet/micro-solutions library

>>105980251
But it ain't JS. It's always some kind of incompatible, transpiled, non-standards extension of JS. Same for this "basedom" thing: It obviously is neither vanilla html nor js. It's a language extension. Not even just a framework.

>>105978995
i swear i replied to this post last thread
>>105937039

>>105977739
this post is a bit schizo but yes wordpress is free and good enough. if you have a VPS already you can look into the nu-PHP runtime with FrankenPHP that can ship a wordpress binary. a bit fucky but then you won't be vendor locked and have to fuck about with random PHP configs

>>105977739
>C gayme dev
I stop reading there. Get a job idiot

The <details> HTML element is pretty based.
https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/Elements/details

>>105980542
Thank you.
>>105981321
That's the idea desu

>>105946460 (OP)
What browsers do you nerds care about hitting?

For me it's everything between chrome and opera released 2 years prior to the current moment. Then chrome/samsung/safari/firefox for mobile. Everything else is a mental illness.

>NOOOOOOOOO YOU NEED TO SUPPORT MY FIREFOX VERSION FROM 2020
Not my fuckin problem. You can pay me if you want my websites to work on your outdated browser and I will personally make it work for you.

>>105982321
opera is chromium now

>>105980343
thanks but I don't really see how that could be seen as commercial use.

>>105980366
nta? what does that mean? I don't understand what this means.

>>105980503
I didn't see how that could be considered for commercial use tho. Also I need something that could work locally, that anyone could download and use. Can that chat app even be done in 100 total hours as a beginner?

I'm building a small webnovel site using only html and minimal java but I'm having problems implementing a search function. I figured I'd start a GitHub and pay someone to finish up my site, any suggestions/advice?

>>105962419
Your thoughts are bloat indeed. They are physical manifestation of the 3D material you sense. Thoughts are bloat.

Soon you just become consciousness that is everything that has a vibration at once. You just are energy.

Is it possible to build a decent calendar in web development if you're a returning beginner and will be spending 100 total hours on it? I need it to have multiple features like 3 or 4 at least.

>>105982321
whatever the tailwind v4 baseline is

What is a good web app SPA that use modern JavaScript tech that isn't made my a mega corporation?

>>105982928
>any suggestions/advice?
copy paste the entire code repo to chatgpt and ask it to implement the search function

>>105983075
Consider making a basic text-only imageboard, will teach you way more than a calendar #9001 project. When I did it it taught me how to work with HTML templates, writing to disk, serving static assets, writing vanilla JS for basic QoL, essential CSS and all the backend like SQL, MVC, auth for mods, etc.
The neat thing is you can make it as complex as you want, so if you don't want to deal with auth yet then write it with no mod tools and basic delete with a password like 4chan

>>105948967
>it doesn't make much business sense to spend time and effort optimising for poor people with old gear.
holy fucking bait

>>105982321
whatever supports CSS nested declarations.
so chrome 130, firefox 132 (although now it works on ESR 128 somehow)

>>105948967
it's usually only kids and women who own high end/the latest phones and laptops
all the well-off men i know how budget 5 year old smartphones and 5 year old $400 laptops, because they aren't fashion items to those people, they have a nice car or motorcycle instead

>>105984965
>5 year old $400 laptops
But it's exactly those devices that work without fuss, my laptop has it's limitations and I understand them and use it solely for work. It chugs along and does it's job well with zero problems unlike some people in my family that need to buy a new one each year (I stopped trying to give advice at this point...)

>>105982321
I care about the current version of Safari that I use
I figure it’s the shittiest browser
I stopped deeply caring about Firefox when Mozilla outed itself as gay as Google
But I don’t really push boundaries, standards-wise for stuff I ship

>>105983955
This
Don’t pay money to a human when you could just get Cursor and spend $100 in Opus 4 credits, the most expensive and best model out there

>>105982847
>i dont want to do any work so you guys must feed me ideas
i hope you are not the standard for new hire because it's fucking disappointing

>>105982321
Error: You need to have the latest chrome nightly to view this site

>M$ shitware eating dirt for abusing shit like Referer [sic] headers
I'm glad to see M$ shit getting hacked for their user agent hostile tracking and retardation. I remember when I had to do stuff like fake user agents just to get M$ sites to offer Kerberos WWW-Authenticate challenges to me as a Linux chad.

Maybe they'll learn something, but I doubt it. They'll probably double down and someone will exploit some other hole.

>>105984935
That's actually exactly what I was looking up, lmao. Nesting is just too good to not use. I cannot go back.

IP over Avian Carriers doesn't have this problem as long as you have enough pidgeons.

I'm a long term NEET looking to land a webdev job, is the odin project a decent start? Already have some experience working with Photoshop if it helps.

>>105989116
if I was just starting I would give up. The job market is fucked beyond repair.

If you still have nothing better to do I would go straight to java and once you land a job (kek) try to branch into something niche like go.

>>105989116
market is fucked and gradually getting worse, >0.000001 job openings for juniors (maybe some internships)
>jeets everywhere
>AI menace growing stronger
>mass layoffs after mass layoffs
these are the odds

>>105989116
>looking to land a webdev job,
I have a CS degree and can't land a job, you're more likely to get a job as a home health aide, they dont have high expectations, all you need is highschool education
https://www.bls.gov/ooh/healthcare/home-health-aides-and-personal-care-aides.htm

Why the fuck do Discord, whether the app or even in just a chrome browser tab, fucking crashes whenever I am playing a game or have Unity3d open? It'll stay open for a bit but then at some point just crash. Sometimes it crashed my entire display and my display goes black for a few seconds before coming back on with everything fucked up. I sometimes even have to close every browser and notepad++ and reopen it cause the windows are in a frozen state.

>>105989116
Consider starting your own company. Building MCP servers is the next big thing: organize some public data and build a prompt library that helps some gen AI model do some task. You then sell access to that MCP server. Lots of money will be made by companies doing this. Teams will shrink because of gen AI, but that also means that you can run fairly sizeable businesses with a small team. We are entering into the golden era of entrepreneurship.

>>105993025
If gen AI is good enough for that, it will also be good enough to build the MCP all by themself thanks to haven that much input for training.

Fuck I wish my mom didn't force a puppy onto me when I'm only starting to learn webdev, it's impossible to fucking focus.
I wish I could just sit in silence and laser focus on it whilst high on vyvanse but the fucking dog won't STFU. I can't even enjoy the ambiance of the rain cause she's too dumb to know she can't go out cause it's raining.
I'll be sitting there finally fixing some dumb thing in React or the backend and she'll fucking whine in a extremely high-pitched decible and fucking kill my vyvanse high and focus and then I just lose all my fucking progress I had made mentally cause I was just taking mental notes.
I've since started taking physical notes but still she won't stop making noise.
This sucks.
You lads who didn't get your progress stripped from you at a crucial time are lucky.
She's held me back a year now, going on two.

What's /wdg/'s favorite CMS?
Today I was listening to one of 404Media's podcast and they mention that the site runs on Ghost which looks nice. Lots of new magazines/newsletters are running on substack but when you go to their site they all look the same.
What do you guys suggest?

>>105948967
you sound like a mediocre dev

>>105948967
Based.

>>105948986
He's correct. Explain to me how poor, incompetent people matter? You can't and I prefer not to associate with you.

>>105984877
>no argument
One concession. Two concession. Three concession four.

>>105993640
>it will also be good enough to build the MCP all by themself thanks to haven that much input for training
I doubt it. There's a lot that goes into getting that data, organizing it, and build a MCP server. Including some subject matter expertise.

>>105995023
>What's /wdg/'s favorite CMS?
the micro-cms I created to update an interactive accordion when I was working and learning jQuery, php and ajax

>>105946460 (OP)
>why your penis should be under 4 inches in size
GOOD MORNING SIRS

>>105948967
They will get worse SEO though and that matters for your business. Having a good FCP is good for business.

>>105997124
google dictates too much and has shittified the web world too much, reminder that angular was the first of the bunch, seo was a scam, and lighthouse performance stats are meant to entretain illiterate tech managers with autistic numbers

the last one bit is not as true, there might be some truth to performance/conversion rate, but the others are painfully true

>>105997150
Speed is one of the most important metrics when it comes to making a web app or website that's fun to browse. It's noticeable even if it's just a couple of 100 ms. It's crazy how little focus there is on raw performance in web dev.

>>105997206
manager type fap to it

>>105997345
*manager types

so whats your comfy, i dont wanna change, stack?

>>105997762
GOAT

GO And Typescript

>>105997762
just building small tools with node, comfy as a motherfuck

>>105998006
>>105997817

webdev thread btw

>>105998288
I'd rather build small tools than a website, a website has way too much complexity, small tools just do the thing
I'll confess that I'm scared of auth

>>105997817
more like
>GOATSE
>GO And Typescript Suck Enormously
>>105997762
html, CGI, bash/Perl, sqlite
It's got a catchy name, too
>esqueseejeepeeelemel-stack

>>105998288
Go on the back-end, TypeScript on the front-end. GO And Typescript. GOAT stack.

>>105998327
Go and TS are both good languages. Easy to use. Easy to make good websites. You can't go wrong.

>>105998515

browsers dont support typescript so you use vite or what? are you using react? what about hosting and servers are you cloudpilled or metalpilled?

>>105998515
TS is a shitshow. It gives people the license to do absolutely dumb stuff. TS is the only "typed" language I know of that does not enforce actual type invariances correct, probably because it neither is an actual superset of JS, nor a statically typed, nor a compiled language. It's bullshit through and through and people (at least speaking enterprise world here).

It's a linter if anything and should only be used as such.

>>105998777
>and people (at least speaking enterprise world here).
Goddamn. People abuse all the shit that it offers and end up creating even bigger piles of shit than before.

>>105997762
bun + tanstack start

>>105997762
Java, Spring Boot, Kafka, ElasticSearch, AWS

>>105997762
Go on the backend. HTML, vanilla CSS (always in style tags), vanilla JS (always in script tags), and fixi.js (reduces the amount of JS I have to write) on the frontend. I also make sure that every initial request is less than 14kB after Brotli (including fixi.js, even though it will get cached), so that the FCP always will be great.

>>105997762
>>105999113
Also, I always use SQLite unless I'm building something very write-heavy. I use Postgres if I'm not using SQLite.

>>105998327
you'd have gotten more style points from me if PG instead of SQLite, but I get the concept of simplicity

>>105998777
>>105998785
I think it's meant for:
>I designed and implemented this
>you shitters maintain it
>there's TS to keep your shitcode incheck to certain degree

>>105999113
how often have you made a real worl application following these rules? sounds like fantasy land 2bh, aka just memeing the current OP

>>105999351
>sounds like fantasy land 2bh
it always is. htmxtrannies are unemployable retards who have never made anything with real users

>>105999388
no need to be so harsh, but I agree

>>105999351
I'm building production apps with it all the time. It's easy: just make sure that the initial request is less than 14kB after Brotli and then you do lazy loading for the rest.

>>105999388
You don't know what you're talking about. Get a job.

>>105982321
Opera 12

>>105999710
>I'm building production apps with it all the time.
can you prove it?

>>105997762
.NET MVC

>>105946460 (OP)
>Websites should be under 14
just like girlfriends

>>105999761
Yes, but I'm not going to. Get a job.

>>105999771
why good morning sir!

could you tell me more about how you ended up there?

>>105999351
....posts like this are just sad, makes you realize the absolute state of web dev. No wonder AI is taking all your jobs.

>>105999794
funny, specially considering that I actually have a job

>>105999920
Grim if true.

>>105999985
at least I don't have to LARP as using the meme stack of the microsecond in real world applications all the time

>>105998630
>vite
Yes
>react
Yes but the appeal of the GOAT stack is that you can use whatever front-end framework you want
>hosting and servers
I have some stuff on free cloud shit but I should probably rent a VPS

>>105998777
>>105998785
I think it's good. If you wonder what type of data a function is expecting, it tells you. And it allows you to catch common errors before runtime (not all errors, but some errors).

react
node + express.js
railway

I call it the NERN stack. simple as.

>>106000031
>larping
Do you even have a CS degree? Assuming that you actually have a job. This is what you're reaching for if you want to build performant apps in production. At least use Preact if you're a Reactmonke.

>>106000060
I think you mean RERN

>>105946460 (OP)
My font alone is 800 kB. Cry about it

>>106000073
im not anon but if you have a cs degree and ended up in web dev.. what went so horribly wrong?

>>106000073
>LARPing
yes using the meme tech take of the microsecond "all the time" "in real world applications", if that's not LARPing then what is it?
where's proof? here it is 3/10 for the pity-replies

>>106000086
I kek'd

>>106000086
You need a CS degree to be a good web dev. That's clear from all the posts here. Lots of people are ok with "good enough" and that's why we end up with slow and bloated web apps.

>>106000103
I guess the CS devs got too lazy huh? shame

>>106000103
Speaking of bloated web apps, new Reddit is so fucking bloated and slow compared to old Reddit.

>>106000125
its because you are being tracked and profiled in real time by government agencies

>>106000125
I remember back when redditors were acting smug on 4chins, laughing all the way to the slaughterhouse

>>106000183
Meds

>>106000242
I sometimes look at the popular posts on Reddit but I use old.reddit.com because it's so much faster

>>106000125
Very few people in web dev care about performance. It's sad but true, just look at many of the replies to this thread.

>>106000458
it depends on who's manages the company, if they care about performance, it will be taken care of, if it doesn't then it'll be overlooked, but when the trend is to hire indians that should give you a hint on just how much they care

>>106000429
the only difference between schitzo thoughts and the truth is a year

>>105946460 (OP)
>14kB
average website today needs 14MB before first contentful paint.

>>106000906

internet connections are faster than ever. hardware is extremely fast. the only reason there are so many dorks trying to optimise this shit is they're paying out the ass for overpriced cloudshit. back to metal boxes. simple as.

>>106000867
datamining the fuck out of everyone and everything is the norm, no longer a conspiracy, prism isn't a secret either

>>106000923
Don't matter if you want a good first contentful paint though and that shit matters for seo. You gon be losing money if you don't have a good first contentful paint. simple as.

>>106001030
>You gon be losing money if you don't have a good first contentful paint. simple as.
that's what google claims, what manager types buy, and it kind of makes sense, but I wonder to what degree though, it goes against what the same google laid out with angular and others like react and vue followed with, however, I'd say front end side of things is kind of simple, optimizing BE integration, which is where the most critical load times (DB) happen is where you have to do some egg juggling
take it with a grain of salt, since I'm not a pro, but is what I've been seeing and taking note of, some companies have an absolute huge monster and clusterfuck in the BE, which "does the thing" but one wonders how can even function with the infinite amount of (maybe unnecessary) complexity it deals with

>>106001119
they're overhyping the importance of these milliseconds for consumers with fake studies. they're pummelling you in the seo rankings because you refuse to buy their solution from them. like the mob guys destroying your shop unless you pay into their protection racket

>>106001314
>like the mob guys destroying your shop unless you pay into their protection racket
sadly that is way too accurate definition of google

>>106001314
>they're overhyping the importance of these milliseconds for consumers with fake studies
Cope.
>Walmart found that for every 1 second improvement in page load time, conversions increased by 2%
>COOK increased conversions by 7% by reducing page load time by 0.85 seconds
>Mobify found that each 100ms improvement in their homepage's load time resulted in a 1.11% increase in conversion
https://www.cloudflare.com/learning/performance/more/website-performance-conversion-rates/

>>106001708
those figures are not too good 2bh

>>106000040
Have you ever tried *not* writing TS and instead used JS with sane datastructures and well defined IO boundaries? It's less code, types are still inferred and it does not need a buildstep. As a bonus: Using JSDoc for typing where absolutely needed will REDUCE the number of times you do that because you will start thinking about shit.

>>106001824
i am conditioned to assume people like you who make sense in the webdev scene are unemployed

ok but how BIG can I make a website before the browser fully craps out?
if i have a webpage with 5 billion table cells with lorem ipsum will that do it?

>>105947585
>I don't know what the QUIC authors were smoking when they recommended the 14 kB rule.
Maybe they just told a white lie to try to debloat the web?

Is this the thread to ask for help setting up a web server at home? I seem to be failing at the networking level and I can't SSH into my machine :(

>>106003274
I think /hsg/ might be more up that alley. I had that issue as well, I set up everything perfectly, but my ISP blocks any traffic that isn't normalfaggotry SNS, netflix and porn.

For the first time in 2 years I applied for a new job. What the fucks with these dumb hackerrank tests? I aced it and then they told me that they weren't interested anymore, but that they had other candidates. Really weird.

>>106003274
i assume you're talking about from outside your own network? ignoring that forwarding your ssh port is retarded and you should be using a vpn instead, are you perhaps behind cgnat? if you are, you're going to need a vpn setup anyways

>>106003483
sounds like you didn't fit the "culture" hard enough

>>106003297
sounds like a problem for tunnels
>>106003274
sounds like a port issue and/or wrong target address

>>106002920
There's a site called Duke Ellington When and Where that has a vast amount of info in tables on a few pages and it used to be super hard to find in page the stuff I wanted but the galaxy brains who make chrome just added more power and now it loads fine

>>106003840
I never even spoke with them. They emailed me that they were interested, so I emailed them back, and they sent me the test, I did it, and then this.

>>106003483
>>106004269
was the task some shit commonly used recruitment tasks or was it something specific like it might been actual problem they were solving in the code kek

>>106005403
It was just simple coding, recursive function, memorization function and a prefix sum function. Major company as well so I don't think they'd try to do that.

What's the best dev environment to work on a twitter-like for fun? Is docker any good/worthwhile or should I just program it normally. I've never done a proper web project from the ground up so even beginner advice is welcome.

>>106003274
it should have been

>>106006610
rawdog your shit on a vps

>>106003483
>got a test
>told it's the best test of all the candidates
>interview with the team
>they say they're not interested
>they say my test had some flaws
well, at least they told me the flaws

big companies play weird mind games, but they also tend to pay well and have desirable working conditions

>>106006674
>>106003483
i dont think this has ever been formalized, but i like to think of it as the "walmart test". shitholes like walmart (note: retail side) will intentionally disqualify applicants on grounds of them scoring "too high" on key metrics like intelligence. this is because they absolutely do not want people who are capable of unionization or critical independent thought in certain positions within their heirarchy. the same goes for any large company, where they actively evaluate the potential threat an applicant poses to the organization in a position. intelligent people are dangerous to have in low-on-the-pole positions because they may break the status quo or expose the company to risk for their practices with this demographic of employee.

punching above your weight is a threat.

>>106003129
Well it could be true but 14 kB is very limiting anyway.
As an example, 4chan banner images are 30 kB on average, these alone far exceed the recommended size.
Obviously I am trying to make my websites as lean as possible, but my target goal is being under 150 kB, not 14 kB.

>>106001824
I just use TS because it's easy.
>JSDoc
Too much typing (by which I mean keystrokes)

>>106006959
The initial request should always be 14kB after compression (gzip, brotli...). You'll get a great fcp if you do this; your website or web app will feel really fast and it's good for SEO. You then lazy load the rest.

>>106006674
>>106006716
Well I was applying for a staff engineer role, it's just weird that they send you an automated message with no feedback at all, telling you they'd like you to apply again in the future. It's just strange. I sent them an email requesting feedback but they just seem to have ignored it. Kind of weird for a major company.

>>106007302
you never interacted with a human. most companies have been outsourcing their hiring pipelines to "ai" comapnies for the last 8+ years now. the only way to get a real job is to network or have an outstanding online presense that grabs the attention of the hn/lobsters crowds.

>>106002014
This unemployed meme is really overused. Yes, I am indeed employed in a company doing extensive web dev. We use frameworks, typescript, buildtools, all the freaking shit.

The code absolutely sucks and _does not scale_ and the reason for that is the fact that people now rely on tools instead of fixing the shit. They fall for the hype and lies that youtubers and other coding influencers tell them without seeing the bigger picture that is the fact that (granted, this may only apply to companies like mine...) are not creating software for themselves to use and care for, but for customers as specified by customers. Which they then will have to support.

So in the end every single dev does not really care, because they are neither really interested, nor are they doing it for their own good. The result of that is a big pile of shit that is absolutely _licensed_ to be that way exactly *because* they are using all the fancy shit so by definition IT MUST BE good, well scaling and maintainable. Right? Right!?!

>>106006610
>for fun
Docker and microservices are the opposite of fun.
For "fun" you should keep it simple and monolithic.
Maybe there's a good reason to break something out into a different service.
Maybe you'll break under the architectural weight of k3s and decide to do something else with your valuable free time.

I learned all of that shit on the job.
In today's market you should prioritize on delivering value to end-users rather than building tech castles in sand.

>>106007312
Well I'm not sure, because at first they just sent me the email that they were interested, but then I replied that before I started taking their tests that I'd first like to have a ballpark figure for what they were expecting to pay me, and then they replied I had to take the test first and that they would discuss that afterwards with me. So there must be some human involved, because that seemed like a handwritten response, at least compared to the previous.

>>106007123
>I just use TS because it's easy.
And that is the key argument here. You think its easy, but easy does not mean producing good results. What you think of as easy is actually just you opting for the least amount of resistance in the short term only.

Good for you if it really works for you. My experience, from my company having the same mindset is the opposite. Once you start working in groups of teams where you cannot - in a friendly way - tell someone in person anymore that they fuck things up, the path of least resistance leads to chaos.>>106007123
>Too much typing (by which I mean keystrokes)
Exactly, so by using it you will start reducing types. Of course you still gotta have the idea of having everyhing neat and clean, but because of that, you will start thinking instead and therefore you will create better data structures, better architectures, less code and better software.

Of course this is only my POV. But this theory should just be as valid as all the other dumb theories surrounding this topic, simply because it just based as much on anecdotal evidence as is everything else in the world of web dev.

>>106007392
>Docker and microservices are the opposite of fun.
imagine being this filtered

building microservices is the most fun I've had in my "career", and I did monoliths and frontend in the past too

>>106007443
I've made entire packaged applications in kubernetes operators.
The conceptual overhead of microservices is insane and 90% of the time you're not doing it because you want independent scaling, you're just mirroring your org chart and are too lazy to get your tech leads and architects to actually talk to each other and build something coherent instead of listening to the siren call of silos

I live in Germany, and I can tell you: it does not matter if your carrier is still on 3G. It does not matter if you have 4G. It does not matter if you have 5G. Internet access sucks in many places, even with a full connection. Latency is high, throughput is low, and your freaking phone will simply lie to you and show full 4 bars.
We still have many people on DSL - not because they prefer it, but because it's the only way to access the internet in many areas.
Hell, the internet access in freaking Croatia is better, even if you're on a mountain with no other person nearby.
Mostly rural areas in Germany suffer from this, but considering that half of the German population lives in smaller cities...

>>106007524 meant for >>106001314

>>106007411
>they just sent me the email that they were interested, but then I replied that before I started taking their tests that I'd first like to have a ballpark figure for what they were expecting to pay me, and then they replied I had to take the test first and that they would discuss that afterwards with me
>>>>>>>>>>>>>>>>>>So there must be some human involved, because that seemed like a handwritten response, at least compared to the previous.

buddy...

>>105999710
you can't provide something useful within 14 thousand bytes /after/ compression?

>>106007139
>The initial request should always be 14kB after compression
Only for HTTP/1.1 and HTTP/2, have you read my post?

>>106006716
in my case I think it was that I didn't "click" with them in the team interview, good, the team was full of roasties and a onions boy lead
regardless, I have to do some research on which wealthy multis are worth applying for in my area, I like that multi, pays well, is near my house, nearly no in-office days, etc.
I just want the bag and remote at this point.

>>106008160
>Similarly to HTTP/2, there is a notion that HTTP/3 and QUIC will do away with the 14kB rule — this is not true. QUIC recommends the same 14kB rule.
https://datatracker.ietf.org/doc/id/draft-ietf-quic-recovery-26.html

>>106008383
>>105947585

Come here once every couple of years, posted this >>106003483 but it feels like most of the people here aren't really seasoned programmers? Not trying to offend anyone, but most of the forums I've visited through 20 years of programming, the developers are quite blunt and direct. This feels more like jeets on LinkedIn. Most of the time I'm just building open source shit, and I'm not active on forums that much. Has anyone ever gotten good advice here?

>>106009117
>Has anyone ever gotten good advice here?
This place is riddled with retarded contrarians muddying the waters, so probably not.

>>106009117
yes I have a few times when I needed it. haven't asked for advice in a long while though

>>106009117
I can call you a nigger if you'd like.

>>106009117
if anything you post is like a generic linkedin jeet post
>I got le ignored again
>so hecking weird amirite guys
not sure what you expected, everyone should know that hiring processes in tech are fucked now

and yes this general especially is filled with js tards, I recommend /twg/ for actual advice

>>106009622
Well like I said, I was doing projects by myself, selling them and stuff. All of a sudden this hiring process became so jewy. You can suck my nuts. In 15 years, I got rejected maybe 2 times, and whenever I got invited to an interview, I never haven't been hired. It's jeets like you shitting up this process.

>>106009660
>has 15 YoE
>can't get a codemonkey job
something tells me that PEBKAC ;)

>>106009689
I only work remote.

>>106009660
nta, yes the general is fucked, the industry is fucked, no need to throw further shit at us though