Ah ok you’re comment was descriptive… while mine was idealistic.

that is not true. Speed and turnaround NEVER wins over skill and quality. You need skill to produce stuff fast that is also of value.

I guess I realise now that the value of something is not what people believe it to be. It is the length of suffering and effort the creator went through.

they don’t solve them. They make it harder for you to make mistakes. Doesn’t stop a capable developer from still introducing vulnerabilities.

that html page without js and just links is MORE secure than anything with JS.

sadly, most likely yes.

Nah… assistants were able to do this before LLMs

Because that other stuff used to work before the company fired all of QA and local devs. Then the bugs after that were used to justify LLMs to replace the “shitty coders” that outsourcing to a sweatshop usually entails. At least with the sweatshops there was some argument to be made that the people working there either had no other choice, or a slim chance they actually cared about their output and made it so at least it would do the bare minimum.

Now that corporate wants to justify their hype and investment in AI to attract the moneyed entities, they will go to any lengths to show it actually works. Even if the Emperor has no clothes on!

The problem with your argument is that you are phrasing that as a problem with how the OS is not able to do what you want. But Linux is able to do whatever you ask it to. The real problem is companies.

Most of the problems Windows users have with Linux is “Software X is not working in Linux” followed by “Alternative Software Y is too weird/quirky/broken on Linux”. This used to be a problem with Gaming. With the investment of Valve into Linux, the scene there has dramatically shifted. Yet, you have cases like that of Roblox whose software is clearly capable of running on Linux but they deliberately hobble it and only support Windows. The important thing is that free software is written and maintained by people in their free time for free. So you can’t expect the same level of polish as a dedicated company working on the software (Of course I can point to beautiful exceptions like Blender, VLC, etc.)

So essentially the problem is two fold:

1. Software/Game vendors don’t want to support Linux
2. Microsoft benefits from having it this way so they bribe their way into having Windows on retail hardware.

Nowadays you can find laptops from manufacturers like Tuxedo or Framework, or even Dell/Lenovo where if you chose to go without windows they often discount your purchase by $100 or in some cases even $200!

So it turns out Microsoft got greedy and is charging like 10% of hardware price as the cost of having Windows pre-installed. (Citation needed, I learnt it here on the fediverse)

You and other people who want their stuff to just work are correct about the assessment of what needs to happen in Linux for it to catch up with Mac or Windows, but are incorrectly attributing the steep gradient set by Microsoft/Apple to inadequacy on the part of Linux.

The incentives are for scammers.

use

display: flex flex-direction: column align-items: center

on the parent container

Stackoverflow is for senior devs to clown on junior devs. It’s the inverse of helping juniors.

I can’t cater to everyone’s needs.

Can’t wait for the law to slap you in the face someday.

oh nevermind… carry on then (Tis a joke about not wanting to touch THAT mess, in case it’s not clear)

I don’t know the details of the MPK. So I consider it as some kind of function that maps {process PID, DLL} => Set of UID. And by UID, I AM talking about the system level user ID. Remember that this feature is a processor level feature. So it has to be transparent to the OS (well at least, to the OS Scheduler). Hence the output of this feature should be understandable to the OS kernel. Or so I hope as the implementation details are vague till now.

which office are you? I am missing such fresh perspectives. Everybody wants someone with 10 years experience in 5 different stacks on 5 different languages/tools.

The company probably wanted a savant for monkey’s pay.

The gist is that a system call is introduced to go into the PCB and change the Effective UID of a process. Security is ensured by a processor MPK which is a CPU provided guard so that a {Process, Library} has only a restricted set of Effective UIDs it can switch to. This operations is supposed to use 30 to 50 clock cycles. So entry + exit is supposed to be done in 100 cycles. This is considered low overhead context switch compared to the traditional context switch on Linux for slower IPCs. They don’t do a comparison against iouring, or simply multi-threaded process.

well yes… that’s what one expects when they go DIY.

or simply use a well designed programming language.

lol this whole article and this thread is a bunch of bikeshedding I haven’t had the privilege of enduring in some weeks.