Hacker News Reader

Comment by simias | Hacker News Reader

I think it's the glue that keeps up with the CPU power curve, so to speak. You give devs more RAM and more cycles and they'll find a way to use them with inefficient languages, suboptimal code and shiny UI.

I think it's important to remember that for instance Zelda: Link's Awakening was less than 512KiB in size and ran on a primitive 1MHz CPU.

But at the same time we have to acknowledge how better it can be to develop for a modern target. We can decide to waste this potential with bloated electron runtimes, but we can also leverage it to make things we thought impossible before.

View on HN

I don't think this is a good example. Zelda: Link's Awakening, while ground breaking, can be, and has been improved upon. Using those resources to create a better user experience is not inefficiency. That's the entire point of a video game - a good user experience.

I think you're not disagreeing with the parent.

You can have an exceptional experience with a game in less than the storage and processing power it takes to run Hello World today;

To that end you can make things better and nicer than Links Awakening with modern resources.

But can you make something 500-3000x better? (that's the order of magnitude we're talking about, with Slack vs Zelda)

No, because improvements don't scale that way.

320x200 was more or less where graphics started. 640x480 was much better. 720p was a lot better than broadcast TV. 1080p was a very nice improvement for a lot of uses. 4K is visibly better, but mostly eye candy. 8K is pointless for a desktop monitor.

The steps give you 4X the number of pixels. At 320x200 you're severely constrained in detail. You'd have a hard time for instance accurately simulating an aircraft's cockpit -- you don't have enough pixels to work with.

1080p to 4K is in most cases not the same kind of difference. There's little for which 1080p is actually constrained in some way and require sacrificing something or working around the lack of pixels.

This isn't because modern software design is bloated and sucks, but simply because improvements diminish. Our needs, abilities and physical senses are finite.

While I agree that returns diminish, I’ll be the first to disagree that those enhancements are unnecessary.

If I do text work on a 1440p 27” monitor I can see the pixels. Anything less than 4K is unacceptable to me. I could see an 8K monitor being a worthy upgrade if I could afford it and my computer could drive it, especially if it meant that I could make the screen larger or wider and still avoid seeing pixels.

Also, we might consider that digital tech is arguably still catching up in some aspects to our best analog technology like 70mm film.

What's wrong with "seeing pixels" when all you're doing is basic text work? Bitmap fonts - or properly-hinted TrueType/OpenType fonts, now that the relevant patents have expired - can be pixel-perfect and are far from unreadable.

Because I have to look at it all day and I’d like it to look smooth.

Also, I can read smaller text when the resolution is higher.

When it comes to technology and innovation I believe that asking “why do I need this?” can often be the wrong question. It feels like a bias toward complacency.

Why drive a car when you’ve already got a horse?

For VR to ever begin to approach something close to reality, and to really prevent people from getting sick wearing the headsets, it needs to be nearly 90Hz per eye and very high definition in a very small, lightweight package. While you're focused on the rectangular screens, the market is figuring out other ways to use this technology.

Yes, VR goes much higher need-wise, but it has a similar curve.

The DK1 was a proof of concept. The DK2 was barely usable for most uses, and not at all for others. The CV1 was about the minimum needed for viable commercial use -- earlier ones needed huge fonts, so many games' UI wouldn't translate well.

By Quest 2 the resolution is okay. It could be better still, but it's no longer horribly limiting for most purposes.

You are technically correct but it might be wise not to omit that 8k on a desktop monitor may attribute to the presented gamut. ( Despite the (sub) pixels not being visible by the human eye. )

> but we can also leverage it to make things we thought impossible before.

Or realistically, to lower hiring costs.

Or shorten dev time, sure. I don't think those are necessarily bad things. The industry seems to have enough trouble as it is to recruit software devs. Can you imagine if you needed deep assembler mastery in order to develop a website? It would have hampered the development of the industry tremendously.

Joke aside, imagine a language that is optimized, to allow for ease of later optimization. As in object orientated, fancy beginner level code goes in, and the hot-loop could be rewritten if needs be by a professional, without doing a complete data-orientated refactoring of the original code.

I don't think that is possible to be honest. I'd love it, but most sub optimal choices require extensive refactorings to work around. I'd love to be proven wrong though.

The current frontiers are nowhere near what is possible though; E.g., python, bash, ruby could be a lot faster if their development was subsidized by governments. Cpp doesn’t have a good package manager (and likely will not have one in the coming decade either). Go doesn’t have obvious features such as string interpolation and generics. Rust’s tooling still isn’t pleasant. ...

ADA and in a minor extent COBOL, were heavily subsidized by the government. Probably even PL/1

From my experience that's a bit like what Cpython is

It would be a set of two languages. One classic, defining operations, sets, flow direction. Then another, which defines containers and the (parallel) paths of execution.

Games are simultaneously more expensive, and cheaper, than ever to make.

At the top end, tech advancements go towards things that couldn't be done before. And at the bottom end, it's enabling smaller shops to do things they couldn't do before because it's now within their budget.

> You give devs more RAM and more cycles and they'll find a way to use them with inefficient languages, suboptimal code and shiny UI.

I'd go as far as saying that most devs do not want to use inefficient languages, write suboptimal code or program that new shiny UI. Quite to the contrary, but users and management demand most of those things just with other names.

really depends on the context. Embedded software has its own fun challenges, but for a simple web app, the last thing I want to worry about is controlling allocations

My boss would never allow me to optimize anything, the only thing that matters to him is write the fastest you can and just make it work and make the client suffer until the client crys and threat to leave the service

> I think it's important to remember that for instance Zelda: Link's Awakening was less than 512KiB in size and ran on a primitive 1MHz CPU.

The Game Boy CPU ran at 4.19 MHz.

Eh, sort of. It's memory bus runs at 1.024 MHz, and operations take 4 CPU cycle multiples to execute. It's really more like a 1.024Mhz processor with the internals running QDR.

Reply on news.ycombinator.com