Видео

I wanne make microvms for pi hole, adguard and pretty much anything so i can run them on top of vps and in my case it will be xcp ng.

Yeah. You have to be incredibly stupid to argue against the existence of front-end frameworks. You'd have to believe companies just decided to create them and use them for no good reason. It's fun I guess.

nice video! What is this terminal theme ?

I'm quite new to rust, but wouldn't it be possible to use features to conditionally build/install the binary? So one can have a a slim worker/master/client etc. ?

This is cool! keep up the great work!

great, keep going!

Great video dude Have you thought about whether this could work without a master server?

Started! Hopefully it will continue to be updated!

This is right up my alley

laravel <3

Keep rust out of everything i use please.

Thanks from Egypt ♥ Please do one for qeme.

Gemini is the worst. Open AI Codex (github copilot) is not bad. Cody / Claude 3 almost never works. Codeium is a scam.

SQLite in WAL mode provides threadsafe read/write concurrency, commits tens of thousands of write transactions per second to SSD, and scales horizontally. The average web app doesn't need vertical write concurrency of running Oracle on a mainframe.

Why you are taking this guy (Pieter) seriously??? I saw the interview and he said a lot of nonsense. Obviously is not a professional programmer.

Software engineers will not be replaced by AI anytime soon, but software engineers that can't use AI to "augment" their skills will be replaced by developers that can. I've been using Supermaven in my VS code for months and find it to be a massive productivity gain. You still need to know how to write code to be able to prompt the LLM properly and to make use of / fix the responses but it's enough to stop me getting blocked on a problem and keep me in a flow state for way longer than without. It's like having a really good pair programer. Your initial prompt could have been easily improved to get a much better response from the LLM. you could have told it what the starting language was, and given it some context on what you'd like to see in the output. The way you've prompted it, It's not your tech lead and the response can't be judged as such, it's a very junior developer with access to tonnes and tonnes of documentation, no idea what to search for and a propensity to mis-interpret it anyway. I haven't tested this but what if your prompt was: ------- Act as a senior software engineer with 5+ years of writing production rust and kotlin applications. I need your help translating an existing kotlin program into a working rust program. Here's the existing kotlin code: ``` YOUR KOTLIN CODE ``` -------

my experience so far: simple tasks: productivity gain complicated tasks: productivity drain

i am not sure that frameworks are better/faster then vanila js/web components. and lot of react js devs simpliy don't know vanila js

its the sqlite for me too, i used sqlite for the first time for android

dont know if i told you this before but i might be skipping js full stack altogether, where did javascript for the backend originate from?

Hello Dear I hope this message finds you well. I'm Emon, a professional digital marketer with a focus on RUclips video SEO and Google Ads. Are you struggling with subscriber and video views for monetization? I wanted to reach out and share how my expertise can help enhance your RUclips channel's visibility and drive more engagement. I was analysing your RUclips channel

I can't speak for companies that are strictly software creation as a product brining it to market. But an organization that has I.T. as basically support, PURE AGILE WILL NEVER WORK!!!! And this is true for pubicaly traded companies that has to report out and document via AUDITS. Budget have to be approved (Meaning a Blank Check to get it done will NOT be given), so there will need to be an initial estimate, and the appliable documentation MUST be adhered to Audit standards for pubicaly traded companies (Charter, Requirements, Tech Spec, Test Cases, Implementation Plan, etc). I've seen an Agilish approach. Development can break up the work into whatever streams they see fit, work on them as they see fit, but documentation STiLL MUST be FOLLOWED!!!! This is a struggle for development Teams that want to follow the manifesto to the letter. There is a reason for having BA's, & PM's on projects. In Pubicaly traded companies I can tell you with CERTAINTITY it is to follow the audit protocols. And YES companies get audited, internally and exteranlly (by the government externally) and NO Leader wants to open themselves up to a failed audit.

Its so easy to trash another language when all you essentially looking for is doing the same thing only faster. Thats not the case with Rust. Rust is not about making your life easier. Its about performance, productivity, security. And yes, it takes time to learn it.

React and Vue are built to have the application live in the browser, completely without a server, if need be. The same goes for Blazor. Those apps will store data on the client and push to the server when available. It adds lots of logic on the client side and basically doubles every domain logic. Htmx requires the server being always on. Which is okay for most 99% of apps. Htmx features should be Html standard.

👋 Author of the original post here! Glad you liked it! And I'm really glad you took the point to heart: that this is an exploration of an alternative idea and not production-ready Python package. I really enjoyed your video. I won't link it here in my comment so as to not catch spam, but would you mind linking to the original article in your video description? Cheers!

Is there already a DSV crate?

let bob = bob, thank you, my next post 👀 thank you for putting the answer in the description! ASCII control characters as he calls it "DSV"

I dont know rust but will try to implement it in cpp

Great video, love it

HTMX

I teach computer science (at a university) -- I started in 2010 and read this article in 2012. What the article discusses is very much true (regardless of what I, or others do we don't seem to be able to teach about half of the students that want to start CS). But the problem is not with assignment (the test that they propose). This is a hurdle I can always make everyone pass. The real issue is that the failing students cannot 'problem solve'. What that means is they cannot take a problem statement (in English) and turn it into a series of loops and if statements. They seem to fail at going back and forth between different levels of abstraction (English description and code implementation). I can always teach everyone if statements and loops on their own / not in a bigger context. Initially I very much focused on adding more time (an additional intro to CS course), and ensuring that students have a proper mental model of what the computer actually does. This helped about 10% of students, moving the success rate upwards... but no where near the amount I was hoping for.

Server side rendering does not require full page reloads. Partials are virtually equivalent to components, the only difference is where the content rendering takes place. Scalability argument doesn't matter for 80% of developers and this applies more if you want a full SPA experience, you can use a hybrid approach where pages are pages and reactive content is reactive content. I think there's space for more than one style of frontend development, plus, why are all the frontend frameworks adopting SSR if it's not the solution?

If server-side is so great then how come apps are routinely rewritten to use a frontend framework but I've literally never heard of any effort inside any company to rewrite an existing frontend framework based app to be server-side only.

Soon someone will post a video to argue that <?php echo $user; ?> is a great idea and we should all be just using that. Give me a break, where do those people come from?

Fundamentally, the server does not know anything about the state of the UI. It cannot know it. Sometimes, that's okay, because you just send the initial state of the UI and don't care about what happens next. But the moment you try doing something more complicated you will need it and a server-side render architecture is just not right for that. If you chose it, you'll be stuck with that forever, and you'll end up making more and more workarounds to somehow make it work. Wanna have a form on a page and also a paginated list? Well, crap you're out of luck, you now have to do partial page renders and manually patch up the client side. It can be done, but it'll be a mess. And the next programmer will have no clue what is going on. While such things just come out of the box with React etc.

"between 30% and 60% of every university computer science department's intake fail the first programming course" it's almost as if people are not instantly good at a thing they are doing for the first time

Not everyone can learn to code because not everyone is logical . As Joshca Bach says , "The difference between a nerd and a normal person is that a nerd thinks the purpose of communication is to submit their ideas to peer review while a normal person knows the purpose of communication is to signal alignment." Or in other words , normal people virtue signal . Nerds side with the facts and then get into serious trouble when the person who is factually correct is a bad person .

The whole "Learn To Code" thing reeked of "Class Warefare" .

I'm not sure this is relevant, but I think it provides some of my learned insights into stalled or failed learning, beyond the ability to just grasp it at the offset. There are other blockers that can get in the way of understanding, those click moments as you put it. I've been trying to learn to code off and on for years. I've worked in IT for most of my adult life, lastly as an Oracle DBA before leaving the industry 12 years ago. Now, I'm back to learning to code again, this time focusing on C++. However, I've had to understand what fundamental hurdles are preventing my own personal progress. For context: I was diagnosed with ASD in 2019 at the ripe old age of 46, which has been a godsend in helping me understand these hurdles. I think I can break it down into five parts: 1. Purpose for Learning. Not everyone is learning to code because they want a job in programming. This certainly needs to be understood by those who teach and for those who learn for that matter. It impacts the teacher/student relationship. It's a bad start if this understanding is missed. 2. Abstraction is a true Learning and understanding Killer. For me, I need to understand what is happening under the hood. In the absence of 2 + 2 = 4, my head BSODs. This is why I've chosen C++, which I know still has a certain level of abstraction. The BSOD aspect occurs in all aspects of my life, hence my favourite saying, "There's nothing the other side of the equals sign." 3. Project-Based Learning. Learning to do another To-Do list for the eleventymillionth time sucks and completely kills the learning spirit. I plateau in the pain of learning something I don't care about and end up dropping out. You just can't concentrate on something you're not interested in. 4. Return to Learning. I can't tell you the number of times I've relearned variables, data types, loops, etc., after returning to learning again. I still haven't come up with the best way to restart yet. I've started looking at Obsidian as a possible way to document my learning. 5. Life's responsibilities. Yeah, this is a real show stopper that just can't be avoided. Very closely related to point 4 I'm hoping that I will develop a winning formula to get productive.

I find that example question interesting. Answers like a = 30, b = 0 imho are not just incorrect in the context of assignments but could also indicate a fundamental misunderstanding of how digital data and computers work. This kind of misunderstanding reminds me a lot of people that I regularly help with computers. They're smart and they understand my explanations, but they lack the intuition to a point where my help doesn't ever stick in the long term. And it feels like I'm the same with maths. I can learn those abstract concepts - kinda - but it feels like I have to mentally translate things through multiple layers in order to get it, it never feels natural or like a native language.

Nothing can match the component ecosystem of js framework. That thing alone is enough to not wanting to switch to server side rendering.

To be honest, whenever I see all of that recommendations to use HTMX, they always demonstrated it through "A Simple Table Application", or sort of a 90's static developer blogs. And this is a main point of usage of it and I not gonna discourage that, but as a say, if you need to do some more complex UI behavior like seamless transitions between html templates, animations, or just simple UI changes that not account server state - HTMX can't handle that as simple as it does those "complex and gross" web frameworks... For me, HTMX niche always in that place, where Alpine.js and Petite-Vue exist right now, but it never gonna be alternative for something more heavy and complex. To summarize it, I just want to say, that all of that "Fanclub of a new shiny thing" guys, they see HTMX as a hammer for anything right now no matter what...

I really like your educational vids, breaking down theories into core ideas and principles and what they mean, and explaining their pros and cons depending on from which perspective you look at them, ty

Using HTMX and some template engine can fix a lot of backend issues including spaghetti code. Then there is rendring part of the code and core api part that feches data. Also if the request header request Json then the api can return Json instead of Html. That way it works like rest service for mobile etc. Only reason where in client heavy application like games etc. This approach does not work well. But many applications are CRUD applications that can benefit from simplicity of operations. Specially when it come to user permission and roles implementation which backend is more close to in an application.

Spaghetti code?😂 Have you ever used a full stack framework template engine like Jinja2. Every component can now be it’s own html that can extend, include or append to a base base html template (with iterations and conditionals). Spaghetti code is any front end framework boilerplate.

I think the whole discussion comes from the feeling that webdevelopment moves really fast and that there is always new stuff out there being worked on. Andrews view sounds a bit nostalgic, like: "back in the days, just HTML, CSS and JS ... don't need all this fancy new stuff, just plain and simple...". I suppose many people vibe with that, but that's not the whole picture. Sure, for a simple static site, just go with that, but if you want something bigger, you will need a more sophisticated approach, which may be more complicated (which is only natural), but less complicated compared to a big project that doesn't have a system and an idea behind it, so: big project (with a framework) vs. big project (without a framework) I mean a framework is basically the result of optimizing your project for something bigger, that's why they exist.

I need a FE framework that brings food to the table!!!

That is the old way of doing it but he is also not relying on anyone else...npm, etc and not many can deal with that BS! Never had an application that wasn't complex so i prefer frameworks like angular or blazor.

Interesting take! Keeping server and client responsibilities separate has its advantages, but frameworks like Hotwire can simplify development and speed up the initial launch, tackling common issues like authentication and security. Often, it's unclear if extreme scaling will even be necessary in the early stages. I've seen a company struggle with Vue for months before switching to Hotwire, which led to a successful and much quicker implementation. I have been working with this solution for over a year and I have to say that everything is very clear and extensible. Even with thousands of users per day, there are still absolutely no scaling issues. It really depends on the project's specific needs and resources. And I think for most apps it's worth starting with Hotwire or something similar.

Advanced Rust is hard, especially if you are not trying to learn the advanced parts of the language but instead trying to make use of libraries or frameworks that do use the more advanced features. That said, if you took the same attitude with C++ you would probably end up with some very very bad code, but with Rust, if it compiles it usually works and most classes of bugs are simply caught by the compiler. Certainly, Rust was one of the harder languages I have tried (used for about three years, but still cant say I have 'learnt' it). The biggest problem we experienced though was that it is a constantly changing language and that combined with the frameworks we were using being in active development meant that we spent over 50% of our development effort just updating to newer versions of dependencies.