"Nobody uses Microsoft Edge". If you work in corporate you'd be amazed at how popular edge has become. A lot of employees who used to download chrome now use edge instead since it's their default.

"Postgres is not very scalable " "Nobody uses Microsoft Edge" Whichever AI is writing his scripts is spinning some wild shit

Alot more people use edge than you would think, just the power of being default browser like safari

You have zero idea how many people use Microsoft edge being default to Windows lol

0:06 - "Nobody uses Microsoft". Oh dear lord... I would love to have GIFs in KZfaq comments...

Watching this on Microsoft Edge.

• They never said they wanted SSR; that's something you made up • Performance gains are performance gains, wether they're in a weird menu or not. They had some real-world benchmarks showing their use case was slow so they developed a solution that can be applied even in spots where performance is less of a concern. * React isn't the default but it kinda feels like that's where you're going with this (I say as someone who gladly writes stupid stuff in React and SSR/SSGs it). People can use other stuff if it suits their needs-even better if they develop a solution can be applied to other codebases. Microsoft made something cool. They had a few options-have a master React instance somewhere, use React Native, or switch to web components. All three are a major refactor and it looks like MS chose something they think might be useful to other people too. Not something to boo. tl;dr Theo is very React-pilled, read this article, and immediately went into defensive mode.

As an Ubuntu user, I've been using Firefox as my primary browser for the last 15 years, and I've always had a de-googled Chromium build as my secondary browser. In 2022 my workplace switched to using MS Teams and Office 365. Its no surprise those webapps don't work well in Firefox, but I hated having to run Chromium just to use those MS webapps. So I tried Edge on Linux, yes Microsoft actually has a build of Edge for Ubuntu, it boggles the mind. Yes I know its literally just the Chromium engine with a different UI, but its unironically actually good. It has more features, faster and easier to use than Chromium, and I'm now using it almost as much as Firefox. It amazes me that Microsoft can make a better version of Chrome than Google can.

the craziest thing about this is that they ever used react for edge

Back in 1998 we had responsive menus that ran on machines with a few MHz. Now we need super computers to render a menu. Amazing! They shoved some super slow shit around so the user wouldn't notice. No wonder the new start menu and Teams are slow.

Just opened Edge and took a quick look around. It's actually quite impressive now. Integrated ChatGPT, ability to split window in 2, the TTS has way better voices than Chrome with multiple language support, native support for vertical tabs and workspaces seems intriguing. Tbh, I probably wouldn't mind using it. It's come a long way since Internet Explorer.

To be very, very clear, they never, not once, said anything about server components.

"Nobody uses Edge." (sees reference to a system with 8GB) "I forgot what shit devices are out there."

Are you really suggesting they use server components for local UI? I know it's in a browser, but if you don't have Internet, I would think you would still want to be able to see the browser controls. And if the "server" in this case is just some local JavaScript engine / runtime, is that not just a restatement of the issue? What performance beneifits would that have?

I use Edge for Windows because it's the fastest browser for that system and it's less CPU and memory intensive than Chrome.

I'm really having trouble understanding Theo these days. He used to sound like he knew what he was talking about, somewhat? Now he is completely clueless and has bad takes on everything. Is he outsourcing his scripts to chatgpt? Or just intentionally posting wrong info for the engagement? In any case, unsubscribed and removed from youtube recommendations... good luck with your company bro hopefully you have a good CTO that can make the important decisions.

We edging react now ?

Calling a 8GB ram device shitty is pretty wild, the fact that it's like that it's because of bad software only

Theo, the one thing you seem to be missing, is that server side rendering is still rendering on the user's device in this case, as edge is a native app that does not fetch its code from a server to run

It's insane that an OS vendor who releases their own native UI APIs would use sth like React for a major software product. What's next? System preferences in Electron?

I enjoy hearing Theo's take on things, but sometimes his American background might limit his understanding of how things function elsewhere

Chrome is an excellent example of slow C++ code, however, in languages like C++, it is extremely hard to write slow code whereas in JavaScript it's extremely hard to write "fast" code.

How can you so confidently deny their actual measurements?? They never mentioned in the article that they use multiple instances of react. They just say that performance is slow because of common bundles. Where did the "20 instances of react" come from?

Theo seems pretty clueless when it comes to Web Components.

This is the result when DX becomes more important than UX. I see this trend all around.

React conf just showed that Microsoft uses React Native everywhere on Windows. Including the menu

Edge is amazing btw, has a ton of features compared to chrome including grouped vertical tabs, workspaces, split layouts, free GPT-4 chat with context, reader mode and way more. I use the stable for work and beta release for personal stuff (to have completly separate tabs and settings). Also using a link redirection app so different links open in the correct browsers.

Firefox's tab ui is the most well throughout, It doesn't shrink tabs into unclickable small upon opening like 50 tabs ... Chrome It's responsive and you can scroll it

Thank you, MSFT, for making Microsoft Edge faster

9:53 - FWIW: Web components absolutely can be server side rendered. Especially using declarative shadow DOM. The standard is fairly new, but since the browser is known/controlled (it’s Edge, so it’s modern Chromium) it renders just fine server-side and is “hydrated” (so to speak, maybe better said “progressively enhanced”) on page load. The DSD of a web component renders 100% before any JS loads at all and no swapping of HTML has to happen at all, since it stays isolated in the shadow DOM, just mounted ahead of time into the WC w/o having to manually create/attach it on construct()/connectedCallback(). That said, I have no clue what architecture, frameworks or design patterns Edge is (or isn’t) using for web components. Just noting that as far as WC’s are concerned, you can definitely pre-render that stuff these days, especially with WASM frameworks like Enhance.

UI isn't just menus. It's things like breadcrumbs, modals, interactions such as enabling/disabling options from other settings, sortable and filterable tables, etc. See e.g. the ARIA Authoring Practices Guide for a list of common UI/UX patterns and how to make them accessible. This is why XUL had a lot of components -- the controls in html forms are too limiting for an application. If you include things like the web developer tools, there are even more UI/UX controls and patterns.

It's not just about menus, but Edge sidebars, such as "Browser essentials". They're basically independent apps ("microfrontends").

I think this is where Qt and QML does a really good job. They precompile your QML so you don't run into what MS is running into. And it's fast.

Chrome also uses HTML for almost everything, you can view the source for pages like history or settings. They don't use React but Polymer Web Components for their UI.

Fun Fact, Start Menu in Windows 11 is possibly React Native. Even the smart screen is just XML and CSS. source: I reverse engineer windows applications for fun.

I don't really understand the relevance of client-side vs server-side rendering in a *desktop application*. The closest we have to SSR is GeForce Now, and that should never be needed for a web browser.

I started using Edge a year or so ago just to give it a fair shake. It was supposed to be temporary. On Windows, I am still using it today. Pretty much exclusively too. Otherwise, it was primarily Firefox. It performs really well and while I am sure that its tab organizing features are not exclusive to it, they are nifty. But hey, I'm just an end user.

They never said their goal was SSR, sounds more like compile time rendering was the goal.

Ive been using edge for a few years now

Using React is always a bad ideia when you have much better alternatives out there....

Flutter and xamarin is dead then?

Microsoft Edge compared to other browsers consumes the least amount of RAM and this matters if you're running multiple tabs or high RAM tasks or have you have low grade laptop or PC

Man, Edge is very optimized for Windows than Chrome

I would guess they could not use react native like you suggested because chrome builds most of its UIs in web tech and they have to hook into chromium

Remember : "shit device" is an economical and ecological choice that makes sense...

This feels like a hamster wheel first UI development was really complicated then we got components frameworks on desktop development and last we got UI editor for C# and Java and maybe QT? With this you could rapid prototyping and got something running in minutes. Now they redo all this with web stuff on desktop.

@10:15 I think if you use declaritive shadow DOM you can get pretty good server rendering. The one thing you'd have to solve is how to serialize and transfer the server's state to the client, but all the existing (and still relevant) frameworks had to solve that too. edit: yep, if I had watched for 30 more seconds the answer would already have been on screen...

*sigh* People hate Edge because they think it's the new Internet Explorer while it has nothing to do with IE. People keep repeating others dumb opinions without checking for themselves. I use MS Edge as "daily driver" and it's great. In the past I've used Chrome (for years), Opera, Firefox (a lot) and few others less known. Edge is really fast, UI can be customized for minimalism. Also, when browsing on the laptop I noticed that other browsers drain more battery, I get the best power efficiency when using Edge. It has the right amount of customization and options that I need while being lightweight. I don't care it's shipped by Microsoft, I care about its performance and options.

I still have a huge sentimental spot for Firefox though...

I was happier when I used to think that anything that wasn't a website was made with non-web technologies

I have stop using chrome/opera since Chromium Edge release. These are the browser I am using! 1. Arc 2. Firefox 3. Edge (for everything).

I'm using react since 2018. Next.js is making me consider to ditch it and move to something else.

I’m on a Mac and primarily use Safari. If I run into a website that has a compatibility issue, I open Edge - which is the only other browser I have installed. Edge is the browser we push for all our users at the company I work at. That’s where I first started using it and find it to “feel” less bloated than Chrome.

I use edge and it's fine, it's not like chrome any better.

Been a web dev for 12 years and I use Edge primarily nowadays. Does nobody else use Edge?

For some reason the phrase "Rube Goldberg Machine" comes to mind ;) I just wrapped up a 30+ year software development career and I'm happy to say I never got dragged into the whole web development mess. For the past 15 years I've been working deep in the plumbing wringing out all the performance I can get from the business logic layer of a variety of in house apps. That's it's own type of mess of course ;)

Could they have used HTMX to achieve this? What exactly is WebUI2.0 and how is it different from HTMX?

Here is the problem. Smart, old-school devs at MSFT are retired or in management. This leaves this new crop of devs who grew up with JavaScript in charge of framework decisions. This BS is now polluting the ecosystem of products and making sh*tty software even sh*ttier. Not just picking on MSFT. It is everywhere.

You use React to make a single page web app that runs on the client, which helps get around copyright issues using fair use.

"Microsoft's native UI build tools suck!" Literally so good compared to any other native build tools/framework that they are one of the primary reasons that Windows dominated the desktop space, but go off king.

Which framework did they replace with react?

Theo, I don't mean to be that guy about it but Microsoft's UI tooling for Windows has been badass for like the past 10 years, React Native just makes it easier to target cross platform, but they're UI tooling itself especially on Windows 11 is genuinely good.

2 years ago Edge only at 3.3% while Chrome at 64.75%, today Edge 5.17% and Chrome is 65.35%

Edge has the best implementation of Vertical Tabs I have been able to find, I can't go back to Horizontal Tabs. The read aloud features work really well, and there is a huge selection of voices. I found out quite by accident the non-English voices will read English in that language's accent. Really fun to try out.

I hate to make the comments go from 420 to 421 but here we go! Single SPA in my experience simplified the import of shared dependencies. Sounds like they made the same thing but a new markup/internal framework.

I'm a MacOS user. I specifically downloaded Edge because its UI is intuitive and lets me focus on the things I want to do without having to fiddle too much with the UI, like looking for which tab I want to use.

sometimes I wish they're collecting audio so they can learn from the times I'm yelling at the screen

3:03 - I'm so used to taking notes in bullet points that I felt called out

Bro the clickbait thumbnail. I was thinking how casual Microsoft culture become.

Another reason why electron got so popular was due to Qt went to shit licensing for any small dev studio, while electron is free,/mit license for corporate use

dude lives in a bubble

A lot of people use Edge. I don’t know why you don’t come out of react, next and chrome There a whole world outside of these things

Microsoft describe the need of "Qwik framework" almost all of that is already solved.

What a timeline we are in where even the experts get eaten alive in comments because things have gotten so complicated and also are changing at an unnecessarily fast pace. Not to mention state actors trying to stick their noses in as much as possible. Somewhere the notion of efficiency just fell by the wayside.

I am a staunch Firefox user but edge is my secondary browser. I don't trust Google at all and thier anti ad block efforts was the final straw.

What does it even mean to say "Chrome uses C++"... both Chrome and Edge are based on Chromium and share a lot. Both use C++, React, etc.

I switched to Edge from Chrome. Same engine, better at sleeping unused tabs. At least at the time. Maybe Chrome has caught up.

Peopel at my workplace dont even know the difference between chrome and edge. They just use whatever works.

I used edge since 2020, i thought nobody uses it. UNTILLL i started working as a front end dev. and the clients are U.S citizens damn they used Edge than chrome. and never saw a client used Firfox

I don't get why we would need to use a JavaScript Framework to make non webpage UI. Microsoft has their own code platform, if we are making UI in html then make that work with C#/C++ and get rid of the interpreters and virtual runtimes

Over excitement in Software Development Community "This is why server components are great"

At the very limit, I could understand using vanilla JS and HTML to handle some UI stuff because that's effectively "dogfooding" and it's quite acceptable in the context of a browser. React though... I mean, it's Microsoft; just walk down the hall until you spot a disheveled dude wearing a Slayer shirt and ask them to write like four lines of C++ to render the menu.

Xbox development team was one of the reasons why MS started supporting Linux internally.

Interesting. I've worked with HTML since 1999, and some computers have had a GUI since, let's say, the '80s. How is it that building A USER INTERFACE is such a huge problem for everyone, still, 30 to 40 years later? It looks like this wheel is being reinvented exponentially more often every year. In 10 years every human on the planet will be reinventing how UI is designed.

So, it's not that they used React, but how they used it that was bad. Does this sound familiar? I mean, it's the case with anything.

Edge is the most underrated browser.

Why shouldn't one? When it came out, it was approximately 10 times faster than Chrome. (It's chromium anyway, so why the fuzz?)

The paragraph you are reading around 8:00 sums it up why JS should not be used in a desktop app. The desktop is meant for NATIVE code. JS is for the browser. Of course modern hardware is fast. But we still should use it efficiently and not wasting resources and energy unsung script languages.

I have switched from Chrome to Microsoft Edge entirely once I checked all the features it had to offer. I never thought such day would come but it was worth it.

They turned their OS into a browser because nobody would use their browser

Crazy that the improvements they are working on have nothing to do with the HTML/CSS rendering. Just shows that all this UI junk is taking useful resources away from the actual purpose of the application. The analogy might be like building a house but focusing on the led lighting, cupboard door hinges and colour of the interior instead of the foundations, framing and plumbing.

it's a similar problem to what the islands architecture in leptos tries to solve

It seems like no matter how fast computers become with Moore’s law, software gets bloated and impedes on the added gains. Developers need to test their code on limited hardware and network speeds.

There was always a "sales pitch" that the "Web UI" could be the UI for all apps, regardless of whether it was "web" or local. Sounds like Windows is still stuck on that.

Edge still has the 3D inspector. That is why I use it for as a dev.

Note that these changes do nothing to improve the utter accessibility failure for those who rely on assistive technology. The doors to FOSS, i.e. GitHub and SourceForge, now feature front doors lifted 3ft off the ground with no stairs let alone wheelchair ramp, with no knob -- there's a switch for that now, but the instructions tacked to the door are written in invisible ink. React sucks.

I mean… are we not all using Edge?!? It’s chromium, so pretty much all of us using Arc, Chrome and Edge are on the same platform…

I think you made a huge assumption about how web components are generated on the server side on 10:21. You stated the process as if it was react components on the server. There are many ways to do server generated web components ie. Resumability. It's even possible for web components to not re-render anything, but merely make whatever is inside client interactive. Rerendering on web components isn't as magical as react does it, because it's more bare metal.

I use edge as my main browser. Most people don't try it and download chrome but edge has some really nice features.

Their initial approach to Edge reminds me of how they handled “transparency" in PowerPoint. There were no alpha channels, Instead, they knocked out a 1-color background by chopping up an image into one or more separate images for *every row* of pixels, and each strip of pixels would start and stop wherever the background color was. It worked, but was fairly sluggish within PowerPoint. It really became a problem when exporting to PDF, where a single slide could top 100Mb presumably due to headers for each microfile.