psilocervine

but wife city is two words

56k warning


cohost (arknights)
cohost.org/arkmints

cathoderaydude
@cathoderaydude

a thing i constantly think about whenever i use a Big Website is "how the fuck do normal people manage this?" like, i know computer. i know what can go wrong with computer and why. i understand stupid bullshit like "what looks like one website is actually 500 separate nodes, even within a single page, and those nodes can even fucking change while you aren't doing anything and forget internal state"

so, when i'm ordering food and the doordash website forgets that i'm logged in in the middle of adding something to the menu, I understand that i just need to reload the page and hope i land on a node that's synced to the rest. to someone who is not Savvy though, this must look baffling, right? they're on a page that simultaneously says they're logged in and not logged in.

huge corporations must lose thousands of orders a day to confusing broken bullshit like this that's easily avoided by just server side rendering the fucking page, which almost invariably has static content anyway. it's not like ecommerce needs fucking dynamic content 99% of the time! they don't need these huge stupid javascript blobs and ten thousand underlying nodes going through five hundred load balancers!


psilocervine
@psilocervine

I swear to god, doing shit on amazon kinda fucking SUCKS. it is absolutely baffling to do things that are standard on other sites I go to? like if I go to newegg or whatever, if I'm at a product page, I can look around the upper right and see its specific product category and just click there to see a list of similar products and then filter them by what I need

on amazon? this feels like a herculean task

also like... navigating the account specific stuff is always so clunky? you are one of the biggest companies on the fucking planet! why are you like this


You must log in to comment.

in reply to @cathoderaydude's post:

i work on a big website and talking with the frontend developers i couldn't even convince them there was a problem with loading huge amounts of slow javascript that pops in randomly as it loads. "computers are fast, the human eye is slow. they won't even notice", they say, ignoring that the load times are in the seconds

you’re right and you should say it. i constantly see my less tech savvy friends get completely stopped in their tracks by bugs i would work around with some combination of editing the url/html manually, devtools, etc. it’s so demeaning to be forced to use broken forms with no obvious alternatives!

I categorically do not Know Computer, but I hang out with enough people that do, and so every time I encounter problems like this (and I do! all the time!) I always think "this must be even worse if you have some idea of why this is broken".

Having crazy-ass multi-team environment where project managers compete internally in some metrics rat race. (that's for short - sorry for the rant below, i'm kinda oversharing how much i hate "enterprise webdev" bullshit)
You may have a dev team owning one subpage of the whole thing, spending less than 20% of time on actual maintenance, and mostly just implementing one AB test after another to squeeze out increases to something (generally the metric that's a theme for this quarter, because it was promised to investors or however they pick that).

I've been working in such environment and it was an absolute bullshit seeing two different teams implementing nearly the same widget independently in different parts of the website to see which one worked better at getting users to provide details for recommendations engine. Two separate apis, two separate "microservices" made to just return a bit of html for one div, just for the sake of testing a hypothesis on production. Then whatever version wins is shipped as is, with no adjustments, because they're already working on another test.

As a user you enter the page and you're assigned to some variant in tens of AB tests running at once. It's like navigating a maze with shifting walls with no guarantee you can even reach the end. And they're okay if you can't complete your order at this moment because you've been unlucky enough to experience a broken mix of experiment variants.

Meanwhile every part of a page has its own bugs and its own way of asking for the same data, so you may see 4 or more calls to api endpoints to check if user is logged in, which also return inconsistent results (half of the page thinks you're a known user, and the other half doesn't recognise you).
This is an environment where just remaking a single page (e.g home page, or product page, or category page etc.) takes more time than it would take to just design and implement a whole store for small-to-medium business (compared to experiences from another job where we did just that).
Oh, and more than half of the page load time is due to tracking scripts so they can measure results of their experiments. Because that's more important than you having a smooth experience.

It's like a leviathan riddled with parasites - it's too big to die, it just keeps going, feeding both itself and whatever lives inside it. Oh yeah, and it keeps feeding on investor funds.

As long as they deliver whatever they promised to investors (or can bullshit with believable excuses for long enough to get more time and keep being funded) then they're gonna keep on doing that. Because it's not about end users or the product itself, or how it's made - just "keep investors satisfied". They're the customer, not the user (and that's why tracking is usually more important than user experience itself - it doesn't matter if the app was made better for the end user if you can't show to investors that some key metrics increased or whatever).

It doesn't die when the app turns to shit - only when funding dries out. It can go on for pretty long time.

in reply to @psilocervine's post: