I'm confused that so many people on this site seem to be working on ancient weak sauce computers. The difference between a 1mb and 100mb app today is nil. I run Slack, vscode, and tons of other apps. Even if they take few gigs of RAM each, there's 62 more where that came from. They all run just fine.
It's like a forum full of carpenters that are using Black and Decker tools to build houses.
Most of us are running other, more important shit that takes up RAM, like an IDE for my code and DB, plus some shells, text editors, etc. If we can get the less important shit to run faster, it would be a win.
Your desktop isn't magic either. Ever run a benchmark with slack turned on versus off? Try it out sometime. You'll be amazed.
> I run Slack, vscode, and tons of other apps. Even if they take few gigs of RAM each, there's 62 more where that came from
Okay, and if your apps used sane amounts of resources, then you could do that on <$1000 of hardware - or even <$100 on a Pi or used machine. Sure, on a programmer's budget/salary it's okay to buy expensive hardware, but it still adds up, and you shouldn't need it.
> It's like a forum full of carpenters that are using Black and Decker tools to build houses.
In that analogy the end result is the same, but the process to get there more complicated. In reality, the difference is between a house that has stuff like insulation and faucets that don't leak, versus than something slapped together in 5 minutes for $50 that costs 50 times more to live in, for as long as you live in it.
It's less effort for the programmer, more effort for every machine it gets run on, every time it gets run. And that stuff compounds in non-linear ways, too: The more data has to be read from disks, the more useless the disk cache becomes, the more data has to be shuffled around in RAM, the more useless the CPU cache becomes. Using gigantic frameworks for programs that use 0.1% of them, which still set up and maybe even poll for all sorts of stuff, not knowing it's not needed, is not "wrong" (certainly not ever when it's open source), but it's more wasteful for what it achieves.
I just checked, and my Directory Opus with 3 windows open, one of them having 2 panes, and all the custom buttons etc. is taking up 100 MB of memory. The list of features is huge. Another thing I have to think of is audio software. Sure, there's also VST that come with gigantic bitmaps or whatever, but generally, efficiency is king. They know their code doesn't run in a vacuum and they are dealing with customers (even when talking about free software) who pay keen attention to such things, to a degree unheard of and unstrived for here.
Sometimes I wish we could just completely freeze hardware development for 20 or more years (if it wasn't for improvements in energy efficiency) just to be forced to become better at programming again. Just look what in 2015 could be done with a machine from 1981: https://www.youtube.com/watch?v=hNRO7lno_DM ... What could we actually do with our current hardware? Who even knows?
Sure, we could optimize the heck out of everything. But at what cost? There is a reason many developers rail against premature optimization. It would cost much more and development would progress much slower. Hardware is cheap. Developer time, not so much.
Hardware isn't cheap, certainly not multiplied by millions of instances. It's just that you don't have to pay for it. But someone down the road will, be it other people or future generations.
You're also confusing premature optimization with skillfull use of resources. Not driving a car with the handbrake on isn't premature optimization.
It's like a forum full of carpenters that are using Black and Decker tools to build houses.