[tdudhhu]

Not only the user is affected by this.

The difference between a 2MB and a 150KB CSS file can be a lot of bandwidth.

The difference between a bad and good framework can be a lot of CPU power and RAM.

Companies pay for this. But I guess most have no clue that these costs can be reduced.

And some companies just don't care as long as money is coming in.

[supertrope]

A lot of companies don't care about end user performance experience. Companies will burden issued PCs with bloated anti-virus, endpoint monitoring, TLS interception, Microsoft Teams, etc. If there's no explicit responsiveness goal, then performance dies by a thousand cuts.

[pixl97]

>Companies will burden issued PCs with bloated anti-virus,

Ugh, bane of my day job. I work with two companies in particular that have high security requirements in their environments and very similar total workloads with our software. One spends around $250k (ish) a year in self hosting costs, the other over a million to get the same throughput. The less costly one worked with us as a vendor to get anti-virus/endpoint exclusions on the file io intensive part of our application and put anti-virus scanning before that point, then harden those machines in other ways. The other customer is "policy demands we scan everything everywhere and the policy is iron law".

[Nemo_bis]

Worst is, nowadays such bloated "security" software is being forced onto Linux servers too... every time I check why something feels slow, Microsoft Defender is hogging resources.

[jillesvangurp]

It's a numbers game. Mostly the difference doesn't matter at all to the vast majority of users. Optimizing for the bottom 1 or 2 percent that don't have any disposable income to update their phones, or pay for your wonderful products or services is not a big priority. And not all companies have rockstar developers working for them. That's why things like wordpress are so popular.

I actually pulled the plug on a wordpress site for my company last week. We now have a static website. It's a big performance improvement. But the old site was adequate even though it was a bit slow to load. So, nobody really noticed the improvement. Making it faster was never a requirement.

What is worth optimizing for is good SEO. There's of course a correlation between responsiveness and people giving up and abandoning web sites. That's why big e-commerce sites tend to be relatively fast. Because there's a money impact when people leave early.

What I find ironic is that the people complaining about this stuff are mostly relatively well off developers with disposable incomes and decent hardware. If they use crappy/obsolete hardware it's mostly by choice; not necessity. Some people are a bit OCD about performance issues as well. They notice minor stutters that nobody cares about and it ticks them off.

2MB is nothing. I'm saying this as somebody who used cassettes, and later floppy disks with way less capacity. But that's 35 years ago. The only time when this matters to me is when I'm on a train in Germany and my phone is on a really flaky mobile network that barely works. Germany is a bit of a third world country when it comes to mobile connectivity. So, that's annoying. But not really a problem web developers should concern themselves with.

[afavour]

Eh. Cloudfront pricing starts at 8.5c per GB and goes down to 2c. I think you’d struggle to use that pricing as a justification when compared to the software engineer hours required to shrink down a CSS bundle. (don’t get me wrong, 2MB is insane and ought to be a professional embarrassment. But I think you’re going to struggle using bandwidth bills as the reason)

I agree with you about frameworks, though. So much waste in creating everything as (e.g.) a React app when there’s no need. Sadly the industry heavily prioritises developer experience over user experience.

[Valord]

This, although I often feel near modern web frameworks (React, similar) do not provide better developer experience.