It's really sad that in spite of the main glaring problem of Electron apps they didn't as much as even mention any work done on performance optimization in 2.0.0. The changes mentioned are just make up when compared to the real problem.
To be honest, people using Electron are doing it because of huge business / process wins and worse performance is a generally acceptable downside. Performance is improving as Chromium improves, but it isn’t a huge focus in the Electron community.
Really, it's as if all this fundamental computer science, O(n), finding the shortest path, all these efforts in optimization that we do at universities - as if all this just went down the drain and the moment you leave the CS department you say, "Ah, screw all this, I'll just allocate all I can, just in case". This is just so wasteful and it really makes me sad that so few people care.
You get all the tradeoffs of Chrome (e.g. cache everything you come by) outside of Chrome proper.
I'm not sure when this became the new normal, but I was taught the exact opposite. How about trying to squeeze that into 200 MB rather than allocating 4-10 GB just because it happens to be available at that moment?
Is it good if a company employs a 100 people but only 15 of them do work and the rest is idle?
Now, as for using all available memory: that's exactly the reasoning that causes so much problems for users, and that's why after so many years of hardware upgrades our machines still feel sluggish: using all resources for your app, only because they happen to be free at that moment, is not OK.
First of all, they are free now, but they might not be a moment later (because another brilliant app author has the same idea and tries to allocate everything). Second, the resources should be used in a reasonable way. If a piece software doing extremely simple computation is suddenly allocating a few gigs of RAM and takes over all cores, my thought isn't "Oh, what an efficient app, it's using all these resources I don't need", but I wonder what went wrong! What is perfectly OK for, say, a 3D rendering node, is completely wrong for an app that could do the same thing with 10x lower resource usage.
I'm not sure if I am able to explain it in more clear terms.
The thing is native apps can do that, and more in less memory. The baseline here is set by "how much is needed", not "how much memory is available".
>Is it good if a company employs a 100 people but only 15 of them do work and the rest is idle?
Yes, if 15 can do the job. then rest can moved for other profitable purposes, or laid off.
An entry-level computer nowadays has 4GB of memory. Not free memory, total memory. That 4GB has to be shared between the operating system, all running apps, and the disk cache.
> and it should be utilised if it's available instead of being wasted by being unused.
If it's available, any modern operating system will use it for the disk cache, instead of leaving it unused.
It's the OS's job to use the free memory for caching the filesystem. If you have special workloads that benefit from more memory you usually change an entry in a config file to the appropriate value or change a settings in the GUI. That's it. No you don't just try to consume as much memory as possible for no reason. You're messing with the OS's ability to use the memory optimally. It's not wasted, you're the one who's wasting it.
These values may not be accurate because they don't include images but they are still very far away from a blanket statement like "150MB per tab".
Clearly we can see that electron has a lot of room for potential and if they fix these flaws there would be no need to critize it, yet they don't even give the slightest damn about it.