That’s complete nonsense I’m afraid. While abstractions are necessary, the bloat of modern software absolutely isn’t. A lot of the bloat isn’t fundamental, but a result of things growing through accretion, and people papering over legacy designs instead of starting fresh.
The selection pressures of the industry do not favor efficiency. Software developers are able to write inefficient software and rely on hardware getting faster. Meanwhile, hardware manufacturers benefit from bloated software because it creates demand for new hardware.
Phones are a perfect example of this in action. Most of the essential apps on the phone haven’t changed in any significant way in over a decade. Yet, they continue getting less and less performant without any visible benefit for the user. Imagine if instead, hardware stayed the same and people focused on optimizing software to be more efficient over the past decade.
Except it’s not nonsense. I’ve worked in development through both eras. You need to develop in an abstracted way because there are so many variations on hardware to deal with.
There is bloating for sure, and of course. A lot is because it’s usually much better to use an existing library than reinvent the wheel. And the library needs to cover many other use cases than your own. I encountered this myself, where I used a Web library to work with releases on forgejo, had it working generally, but then saw there was a library for it. The boilerplate to make the library work was more than I did to just make the Web requests.
But that’s mostly size. The bloat in terms of speed is mostly in the operating system I think and hardware abstraction. Not libraries by and large.
I’m also going to say legacy systems being papered over doesn’t always make things slower. Where I work, I’ve worked on our legacy system for decades. But on the current product for probably the past 5-10. We still sell both. The legacy system is not the slower system.
If I write software for fixed hardware with my own operating system designed for that fixed hardware and you write software for a generic operating system that can work with many hardware configurations. Mine runs faster every time. Every single time. That doesn’t make either better.
This is my whole point. You cannot compare the apollo software with a program written for a modern system. You just cannot.
I’m not disagreeing that it’s different. It’s a more fair comparison to compare it to embedded software development, where you are writing low level code for a specific piece of hardware.
I’m just saying that abstraction in general is not an excuse for the current state of computer software. Computers are so insanely fast nowadays it makes no sense that Windows file Explorer and other such software can be so sluggish still.
You’re making the fallacy of equating abstractions with inefficiency. Abstractions are indeed useful, and they make it possible to express higher level concepts easily. However, most of inefficiency we have in modern tech stacks doesn’t come from the need for abstraction. It comes from the fact that these stacks evolved over many decades, and things were bolted on as the need arose. This is even a problem at a hardware level now https://queue.acm.org/detail.cfm?id=3212479
The problem isn’t that legacy systems are themselves inefficient, it’s with the fact that things have been bolted on top of them now to do things that were never envisioned originally, to provide backwards compatibility, and so on. Take something like Windows as an example that can still run DOS programs from the 80s. The amount of layers you have in the stack is mind blowing.
That’s complete nonsense I’m afraid. While abstractions are necessary, the bloat of modern software absolutely isn’t. A lot of the bloat isn’t fundamental, but a result of things growing through accretion, and people papering over legacy designs instead of starting fresh.
The selection pressures of the industry do not favor efficiency. Software developers are able to write inefficient software and rely on hardware getting faster. Meanwhile, hardware manufacturers benefit from bloated software because it creates demand for new hardware.
Phones are a perfect example of this in action. Most of the essential apps on the phone haven’t changed in any significant way in over a decade. Yet, they continue getting less and less performant without any visible benefit for the user. Imagine if instead, hardware stayed the same and people focused on optimizing software to be more efficient over the past decade.
Except it’s not nonsense. I’ve worked in development through both eras. You need to develop in an abstracted way because there are so many variations on hardware to deal with.
There is bloating for sure, and of course. A lot is because it’s usually much better to use an existing library than reinvent the wheel. And the library needs to cover many other use cases than your own. I encountered this myself, where I used a Web library to work with releases on forgejo, had it working generally, but then saw there was a library for it. The boilerplate to make the library work was more than I did to just make the Web requests.
But that’s mostly size. The bloat in terms of speed is mostly in the operating system I think and hardware abstraction. Not libraries by and large.
I’m also going to say legacy systems being papered over doesn’t always make things slower. Where I work, I’ve worked on our legacy system for decades. But on the current product for probably the past 5-10. We still sell both. The legacy system is not the slower system.
Abstraction does not have to imply significant performance loss.
It does. It definitely does.
If I write software for fixed hardware with my own operating system designed for that fixed hardware and you write software for a generic operating system that can work with many hardware configurations. Mine runs faster every time. Every single time. That doesn’t make either better.
This is my whole point. You cannot compare the apollo software with a program written for a modern system. You just cannot.
I’m not disagreeing that it’s different. It’s a more fair comparison to compare it to embedded software development, where you are writing low level code for a specific piece of hardware.
I’m just saying that abstraction in general is not an excuse for the current state of computer software. Computers are so insanely fast nowadays it makes no sense that Windows file Explorer and other such software can be so sluggish still.
Dxvk would like a word
You’re making the fallacy of equating abstractions with inefficiency. Abstractions are indeed useful, and they make it possible to express higher level concepts easily. However, most of inefficiency we have in modern tech stacks doesn’t come from the need for abstraction. It comes from the fact that these stacks evolved over many decades, and things were bolted on as the need arose. This is even a problem at a hardware level now https://queue.acm.org/detail.cfm?id=3212479
The problem isn’t that legacy systems are themselves inefficient, it’s with the fact that things have been bolted on top of them now to do things that were never envisioned originally, to provide backwards compatibility, and so on. Take something like Windows as an example that can still run DOS programs from the 80s. The amount of layers you have in the stack is mind blowing.