Love tech?🤩 Hate spam?🙄
Subscribe to our weekly newsletter!
Like what you're reading? Share it!
Follow us to stay up to date with our latest and greatest!
4
minute read

360° IT Check #8 - Vue 3.2 released, Facebook open sources their data center solution, Unity now has a solution to the work-from-home problem, and more!

360° IT Check is a weekly publication where we bring you the latest and greatest in the world of tech. We cover topics like emerging technologies & frameworks, news about innovative startups, and other topics which affect the world of tech directly or indirectly.

Like what you’re reading? Make sure to subscribe to our weekly newsletter! Listen to this episode on Spotify:

Vue 3.2 Released

The newest release of the popular JavaScript library is here with us, and with it there are few promising changes. Let's start with the ultimate feature - speed. The new release is supposed to be faster. Changes are concerning primarily the reactivity system. It uses ~17% less memory, and dependency tracking is ~40% faster. The v-memo directive, introduced in this release, allows Vue to be one of the fastest mainstream front-end JS library.

That’s not all! The improvement to server-side rendering capabilities of the framework is a fact, with the new @vue/server-renderer. It is now decoupled from Node built ins, which allows for running it in serverless scenarios.

Furthermore, all users of Web Components should be happy. The update introduces a way to create custom elements. You are now able to generate whole component libraries which will work with or without any framework. They are a great way of keeping a company design system in sync across many frameworks, and even platforms (Electron).

An important note, is that the code generated by this release’s template compiler will be incompatible with older versions of Vue. This is particularly important for authors of component libraries and such.

For a full list of changes go here.


Facebook’s Time Card

What do you do when your scope is international, and you need to coordinate data across the world with millisecond precision, what do you do? You develop a solution for that. Time Cards, company’s answer are now open-source

The company from Menlo Park, California cannot rely on the existing solution - the Network Time Protocol. It is simply not precise enough for FB’s scale. Margin of error of 10 milliseconds was simply too much - it now is 100 microseconds (1 millisecond = 1000 microseconds = 0.001 second). The company elaborates further on their reasons - they mostly are related to high costs of operation, the reluctance to deal with proprietary software and hardware, and technical shortcomings of existing solutions.

You can build your own time card by following the instructions listed on this GitHub repo.

Unity acquires Parsec

The parent company behind Unity acquired Parsec, the company offering “high performance remote desktop and streaming technology”. This is a dev-facing move, rather than a gamer-facing move - and it might not even be necessarily a purchase to make games created using Unity more impressive. Ever since we left traditional offices, taking office equipment home is not problematic in case of less powerful laptops. When we think about companies creating animations/visualizations/games, however, they have to have much more powerful equipment, which sometimes is too heavy to take home. This is exactly where Parsec comes in with their promise of low latency, ultra high-quality streaming. Can’t take your computer home? No problem!

Even though Unity, and, perhaps its biggest rival, Unreal were traditionally connected with video games, it is not the case anymore. They have become advanced tools for creating films & videos, but also AR/VR solutions, and even architects & engineers can use the engines’ advanced features.

From Debian To Arch - Steam OS

Valve, the company behind the upcoming Steam Deck, introduced a change to their Steam OS, which is going to power the company’s console. It’s a major one as well. The company decided to change the foundation of its operating system from Debian to Arch. The latter is lighter & easier to customize, although that does not seem to be the reason why the company is going for the change.

The change is motivated by the ease & the speed of development. Valve designer, Lawrence Yang explained everything to pcgamer.com:

So, Arch Linux, one of the main reasons, there's a couple, but the main reason is the rolling updates of Arch allows us to have more rapid development for SteamOS 3.0. [...] We were making a bunch of updates and changes to specifically make sure that things work well for Steam deck, and Arch just ended up being a better choice for them.

In short, Arch is developed much differently than Debian. Debian has stability in mind, first and foremost, and they cultivate the “if it ain’t broke, don’t fix it rule.” Arch is based on a “rolling” model. The fact that this distribution is in general more recommended on PCs does not hurt either.

OpenAI Announces Its New Release Of Codex

An improved version of Codex, the natural language processing model from OpenAI, was released in beta last week. This is the same (or a very similar) model that powers GitHub Copilot

The model itself is a descendant of GPT-3 - the very famous model that got lots of people excited about its capabilities. The new release is around 14 KB of Python code - its predecessor was only 4 KB heavy. This results in the new model being able to process 3x the amount of contextual information, company says. It shows for sure - OpenAI demonstrated how one can create a command line tool to rewrite scripts from Python to Ruby or how it’s possible to write a simple game with the help of Codex. If you want a hands-on example from a third party, look no further.

Here we can see a simple header being created


While here we see how a countdown can be implemented in JavaScript, to then use it a WebView-based mobile framework


The model itself won’t replace programmers, however. Not now, and not in the near future. The company clearly aims to lay out foundations for tools aiding them in their everyday work. All developers have mundane tasks they wish they could “outsource” to an AI - say, writing a lot of boilerplate code, for instance. GitHub Copilot is an example of such use of OpenAI’s product, and one that is already aiding creators in their work - not without problems, though.

Categories:
Check Out Our Latest Podcast Episode

Related articles

© 2021 ITMAGINATION. All Rights Reserved. Privacy Policy