I spent some time this morning thinking I was wasting memory using an immutable vector type in c# (RRB tree). It made things a lot easier and probably safer, but it ended up using about 1 MB more memory and "wasting" a couple of hundred ns. Then I read about stuff like this and I feel better about myself.

Because that html page needs a full browser to render and browsers have accumulated decades of bloat for everything from webgpu to drm. There's a reason we can't write new web browser from scratch, and why some people are excited for protocols like gemini that you can hack a user agent over a weekend.

There's a bunch of other requirements you need to meet and no one cares about RAM consumption. So you do it the most time efficient way, not the most memory efficient.

Before you scorn, doing it the opposite way, taking more to use less memory, makes you less promotable, not more.

Perhaps no one cared about RAM consumption, but RAM is expensive again.

In this specific case, is your Weather app important enough to waste 1GB on your users' PCs? Most PCs only have 16 GB.

>> is your Weather app important enough to waste 1GB on your users' PCs?

I agree. I don't want or need a weather app on my machine.

Did not ask for it and I do not use it. Why should it be using any of my RAM?

And this is why Windows will have it's lunch eaten by Linux.

Why exactly? Because all the people who really care about good software will switch over? All 1% of the market?

Most people don't care about most things. I don't think much about the vast majority of products I use, because there's only so many things I can track closely. Same goes for everything else. Most people won't look and won't care.

I've heard SO MANY people who wouldn't usually care about such things say unprompted that they hate Windows 11. The ads, the nagging, the unbelievably slow performance for things that used to be instant, the terrible taskbar, etc etc.

I think they've finally made it so shit that people actually care and will look for alternatives.

Maybe it won't happen today, maybe not in the next 5 years, but it will happen.

I’ve read this sentence for more than a decade now.

Honestly it just reveals a poor theory of mind regarding the average consumer.

I wonder when this praised day will come. When did we hear this first? Is 20 years ago too soon?

When?

Just the absolute laziest implementation

Embedded dev here too. I work in 32KB RAM and even that's massively overspecced for my current usage. I can't imagine what kind of data I'd need to store to fill up a GB

There's probably multiple devs, which is part of the problem.

Ironically AI could likely do better than this if prompted right.

My /dev/random could do better if queried correctly