Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

High-Level Is the Goal

bvisness.me

Post reply on HN
Print view View original

131–140 of 158 posts

Re: High-Level Is the Goal

#131

Post by bvisness »

teiferer wrote: post #67

> a gigantic manual that lists every property of the system in excruciating detail, which is totally worthless for learning and barely usable as reference. It's the only usable form of reference! I want all the details to be presented in a reference. Where else? > low-level tools are terrible too It seems to me the author is confusing lack of familiarity with lack of existence. There are lots of fantastic tools out t…

To be clear, I also appreciate good reference. But reference is bad to learn from, and frankly a big PDF is not even a very usable reference resource when compared to, say, MDN.

As for lack of familiarity, I am certainly not some grizzled 80s Unix veteran, but if you're comparing the Chrome or Firefox dev tools vs. `perf`, it's no contest. Browser dev tools help devs build an accurate mental model of what their software is doing. It is extraordinarily rare to find command-line tools or other low-level tools that are well designed enough to do this. (Maybe Superluminal now? Tracy?)

Re: High-Level Is the Goal

#132

Post by greatgib »

torginus wrote: post #80
cyber_kinetist wrote: post #11

I think the real conclusion is: someone has to make a native cross-platform desktop UI framework that doesn't suck. (Yeah Qt exists, but it really sucks...) Until then, everyone will default to just using the browser for a desktop app, and the beatings will continue. Because of this, I'm really looking forward for PanGUI to step up ( https://www.pangui.io/ ), their UI framework is very promising and I would start usi…

Why do you think Qt sucks? Other than the C++ focus, and basically the framework's decision to make you use everything they give you, which is controversial. But apps made with Qt as an end product don't think suck. Qt is a fully featured and modern and high quality framework.

Apps made with qt are mostly good but the engine sucks hard indeed. The programming is difficult, let's say compared to GTK, the SDK interface change totally every couple of years forcing apps to rewrite everything from scratch. Probably related to that, the documentation is splitted in an hundred places, often undecipherable between what is obsolete or for current version. Most parts of the framework are not really documented anyway, there is just automatically generated reference guide from the code that is missing most part that is needed. Like you have an option name, now what does it mean? What this option to be used for? It's hard to extend simply to bypass default behavior of the SDK that could be limiting. Let me not speak of the whole qml craziness and the qtuieditor... And the worse of it is that the license is (or was because I don't know recent evolution) totally shitty. Like free to use in that case and that case, and commercial in that case and that case. No one serious would like to build something on a private garden like that.

Re: High-Level Is the Goal

#133

Post by bvisness »

analogpixel wrote: post #123

> Truckla is an excellent execution of a flawed idea. If you want to build a good pickup truck, you have to start with the frame. But the "truck" that was "built from the frame" by Tesla was somehow worse. So his argument is frameworks built from the ground up will be ego projects that are big, ugly, shiny, and don't actually work

My whole argument is invalidated because one stupid company designed one really stupid truck? This just in: because pickup trucks have already been invented, all new attempts will result in Cybertrucks.

Re: High-Level Is the Goal

#134

Post by greatgib »

wiseowise wrote: post #62
cyber_kinetist wrote: post #27

Earlier quoted context omitted.

Not bad for mobile apps, but still sucks a lot for desktop support. Also, really wished they've opted for a more general language like C# rather than Dart - but that's inevitable since Google needed to make use of their Dart language after they've failed to standardize it on the Web (and I think they don't want to use a language developed by Microsoft of all companies)

They've picked Dart because it was the only language that could have small aot binaries, hot reload capable runtime without compromise and most importantly because they could influence development of the language. C# is one of the worst choices they could make at the time.

I don't think so, I think that some teams at Google were trying hard to push Dart that no one wanted. And so Flutter was about to create a framework that was supposed to be the main and only one for Android app and co in the future (at that time) to force us on switching to Dart.

Re: High-Level Is the Goal

#135

Post by bvisness »

torginus wrote: post #96

I think people overestimate the necessity of 'high level' conveniences or the difficulty of writing C/C++ to the metal. For example, take Dear.IMGUI which is a c++ UI framework with a kind of data binding, which generates vertex buffers which can be directly uploaded to the GPU and rendered. It supports most of the fancy layout stuff of CSS afaik (flexbox etc), yet its almost as low level as it gets. The code is also…

I think it's easier to write than React! (And remember, I did React professionally for many years.)

People assume that surely the pile of high-level languages and frameworks will actually result in a better experience. But Dear Imgui or Clay are genuinely easier to use, easier to iterate in, and faster too. This is why I think it's so unhelpful to focus on the position in the stack: Dear Imgui is a better high level tool than React, despite being "low level" by modern expectations!

Re: High-Level Is the Goal

#137

Post by cyber_kinetist »

tialaramex wrote: post #127

> If you have heard of the Handmade community, you likely think... I think that "the Handmade community" are people who mistook Casey for a messiah, in a sort of "Life of Brian" scenario. The best part of "Handmade" is when somebody who'd otherwise doubt their ability makes a video game. The resulting games are usually at least interesting, and video games are art so, interesting is enough. If Steam had a dozen inter…

I think as typical niche Internet "communities" do, the most vocal will naturally be snobby elitists who deride at anything that doesn't fit in their narrow worldview, and that mob will naturally represent the "community". (I also think Jon Blow has his part in enabling that, but oh well.)

But I'm still impressed what Casey was able to organize during the past few years, it's clear that what he's published as educational material (Handmade Hero, Computer: Enhance) has invigorated interest in low-level programming. Even though there might not be immediate results, I think this will matter later on when the people who grew up with these materials will now build new infrastructure for the post-Web era.

Re: High-Level Is the Goal

#138

Post by greatgib »

nchmy wrote: post #121

I know this article is largely about too many js frameworks and also a call to making better tooling for low-level programming. But it is also largely about the fact that most people making those js frameworks are not low-level programmers. Folks really ought to check out Datastar. It's a newish js framework that is extremely innovative despite leaning as much as possible on browser apis. It's htmx + alpine + much mo…

Just my point of view but I like the vibe of the htmx creator but think that htmx sucks to use it more than for a demo case.

If the guy was not cool and posting fun memes, no one would even look at that.

You go backward in term of performance by forcing a constant stream of heavy content between the app and the backend, state is hard to manage, The backend would have again to take charge of the CPU intensive part of the app for users. The framework is limiting by trying to be very simple, fitting tags and hiding the complexity to the dev (in an "automagic" way). But that is the worse as, as soon you want to deviate from the main documented highway case, it will become nightmarish.

Re: High-Level Is the Goal

#139

Post by qingcharles »

nchmy wrote: post #121

I know this article is largely about too many js frameworks and also a call to making better tooling for low-level programming. But it is also largely about the fact that most people making those js frameworks are not low-level programmers. Folks really ought to check out Datastar. It's a newish js framework that is extremely innovative despite leaning as much as possible on browser apis. It's htmx + alpine + much mo…

I'm switching to Data Star because it feels like it exists in the golden zone between HTMX and React.

Re: High-Level Is the Goal

#140

Post by nchmy »

greatgib wrote: post #138
nchmy wrote: post #121

I know this article is largely about too many js frameworks and also a call to making better tooling for low-level programming. But it is also largely about the fact that most people making those js frameworks are not low-level programmers. Folks really ought to check out Datastar. It's a newish js framework that is extremely innovative despite leaning as much as possible on browser apis. It's htmx + alpine + much mo…

Just my point of view but I like the vibe of the htmx creator but think that htmx sucks to use it more than for a demo case. If the guy was not cool and posting fun memes, no one would even look at that. You go backward in term of performance by forcing a constant stream of heavy content between the app and the backend, state is hard to manage, The backend would have again to take charge of the CPU intensive part of…

Are you replying to the right comment? Mine was about datastar, not htmx.

Though, you seem to be arguing against hypermedia in general - a stance which is just plain wrong in most cases. Htmx's essays and book do a good job of addressing why.

Post reply on HN