Naomi Pentrel

https://lnkd.in/eaiK35Ku Java 25 is now feature-frozen and on track for release in September! Mark Reinhold announced JDK 25 entered Rampdown Phase One with 18 features locked in – and yes, Java 25 will be a long-term support (LTS) release . Also, planning for Java 26 is already kicking off (the expert group is formed), so the innovation train isn’t slowing down. What’s coming in 25? Some pretty neat stuff: Structured Concurrency (Preview) to simplify multi-threaded code (no more callback hell with threads), Scoped Values for safer thread-local data, Primitive types in patterns (pattern matching getting richer), and experimental JFR profiling improvements for better performance insights . Plus, a bunch of housekeeping like removing the legacy 32-bit x86 support – bye, era of 32-bit Java! For us developers, an upcoming LTS is like a new baseline. Many of us still chilling on JDK 17 or planning a move to 21 – but 25 signals where Java’s headed and what we’ll standardize on for the next few years. If you haven’t tried the previews in 19, 21, etc., now’s a good time to peek: these will likely finalize in 25 or 26. The fact that JDK 26 work is starting means features that don’t make 25 (say, some Project Valhalla value types?) might land in the next cycle. Java is alive and kicking, dropping improvements every six months. For enterprise folks, the rampdown means it’s safe to start testing your apps on JDK 25 builds – see if any deprecations bite, and plan to leverage those new APIs. September will be here soon. Props to the OpenJDK contributors for keeping Java fresh. Happy (early) LTS! #jdk #java #enterprise #programming

222

3 commentaren

Over the past year something’s been quietly simmering in the background. Late-night writing sessions, countless edits, and a lot of coffee. Today I’m excited (and a little proud) to finally share it: 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗣𝗼𝘄𝗲𝗿 𝗕𝗜 𝗩𝗶𝘀𝘂𝗮𝗹 𝗖𝗮𝗹𝗰𝘂𝗹𝗮𝘁𝗶𝗼𝗻𝘀: 𝗦𝗶𝗺𝗽𝗹𝗶𝗳𝘆𝗶𝗻𝗴 𝗗𝗔𝗫 Written together with Jeroen (Jay) ter Heerdt and Madzy Stikkelorum-Groenveld, this book is all about making one of Power BI’s most exciting new capabilities of the past years, 𝗩𝗶𝘀𝘂𝗮𝗹 𝗖𝗮𝗹𝗰𝘂𝗹𝗮𝘁𝗶𝗼𝗻𝘀, clear and approachable. We focused on real examples and practical tips so that anyone, from business users to Power BI experts, can understand how visual calculations work, explore common usage scenarios, and navigate them confidently without getting lost in the complexity of DAX. If you’ve ever wanted to simplify your Power BI models or just explore what visual calculations can do, I think you’ll enjoy it. 👉 It is not available yet, but soon it will be! Pre-order a copy and let us know what you think: https://amzn.eu/d/cmaFvWc #PowerBI #VisualCalculations #DAX #DataAnalytics #BookLaunch

381

66 commentaren

Vibe code is legacy code. (A great article from Steve Krouse mentioned this yesterday - link at bottom). Essentially: • Letting AI crank out code you barely read = tech-debt on fast-forward. • Legacy code → complex, no one understands it, nightmare to maintain. • Vibe code → same outcome, just faster. The key difference is that legacy code is traditionally undocumented code, with untrustworthy technical docs, and zero functional docs. If we are vibe coding like "BUILD ME A TIKTOK CLONE PLZ THX" - then yes, the comparison to legacy code is 100% valid. ...but if we are providing vastly more context and leveraging Spec-driven development (SDD) - then we are actually building the documentation first. And sure, if the code isn't perfect, we tweak the docs, regenerate the code - and voila, we have living documentation. We've drunk the Kool-aid @ Userdoc, but today I'm seeing this from every zen master AI coder. ... I'm not sure if I'm the only one saying "SDD" for spec-driven dev. but I would bet money it will be a common acronym within the next 4 weeks... and in the dictionary by November 😜 Article: "Vibe code is legacy code" - https://lnkd.in/gRihTX75

21

7 commentaren

Most companies think their internal wiki is in good shape. But open it up and: - Half-written guides or at least not up to date - Drafts from 6 months ago - “@someone please update” comments - Broken links Sound familiar? It’s not because people don’t care. It’s because creating & maintaining docs is hard, thankless, and never urgent compared to hitting revenue goals (I too have dealt with problem time and time again). Wikis were supposed to fix this. A single source of truth. But in reality? Most are graveyards of outdated info. Here’s the real problem: 1️⃣ By the time you’ve written & formatted the perfect doc, the process has already changed. 2️⃣ Updating is a heavy, manual lift, screenshots, rewording, linking, formatting. 3️⃣ Most tools make updates just as painful as creating docs from scratch. So teams give up, outsource it, or scramble at the last minute. There’s a better way: Stop writing docs from scratch. Start recording. Record yourself doing the task once, then let tech do the heavy lifting. That’s why I built Zarta, it turns your Looms, Zooms, and screen shares into structured, step-by-step guides with screenshots & hotspots. When things change? Upload a new video and the guide updates. No more wiki graveyards. No more “@someone please fix.” Docs that stay alive while you focus on work that actually drives the business. Link to the blogpost 👇 https://lnkd.in/dTR44JV2

6

🚨 Major ARIA Update on the Horizon! WAI-ARIA 1.3 is moving forward—and with it, a wave of new features that could change how we approach accessibility in custom components. 🗓️ Expected Timeline: A final Candidate Recommendation (CR) draft of ARIA 1.3 is targeted for early July 2025, with the final Recommendation expected shortly after. 📌 So what’s new in ARIA 1.3? This update brings critical improvements for assistive tech support and developer clarity: 🔹 aria-description – Offers a clean way to supplement aria-label with additional details 🔹 New roles – caption, code, deletion, insertion, and more—designed to support richer semantics in content editing and communication 🔹 Improvements for authoring tools – Clarifies usage of roles like application and presentation 🔹 Better support for complex widgets – Enhanced guidance on composite roles and keyboard navigation These changes are subtle but hugely impactful for those of us working in custom UI development. If you're building components beyond native HTML (e.g., in Webflow, React, or Vue), this is a spec to watch. 👀 🎯 My favorite part? The addition of aria-description. As someone who works daily to make complex interfaces screen reader-friendly, this small change gives us cleaner authoring and better UX for assistive tech users. Read the full W3C working draft and stay ahead: 🔗 https://buff.ly/Dme2P3J 💬 Are you excited about ARIA 1.3? Have you run into friction points with ARIA in your work? Let’s talk about it.👇 #A11y #ARIA #WCAG #InclusiveDesign #Accessibility #ScreenReaders #W3C #WebDev #WebAccessibility #SemanticHTML #GracefulWebStudio #DesignForAll #A11yBookClub

209

11 commentaren

🚨 New episode of Señors @ Scale is live! Trailer below 👇 🎙 Observability at Scale with Erik Grijzen, Principal Software Engineer at New Relic This one’s for architects, staff engineers, and anyone scaling complex frontends across multiple teams and products. Erik has been: 🏗 Building one of the first large-scale micro-frontend architectures before the term even existed ⚙️ Leading UI platform teams powering observability for millions of developers 🧭 Guiding technical direction at New Relic through influence, not authority In this episode, we cover: ⚡ How New Relic built its unified UI platform that lets teams ship to production in a day 🔍 Why observability isn’t just a dev tool anymore—it’s a business necessity tied to uptime and revenue 🧩 The challenges of micro-frontends at scale: performance, duplication, and dependency hell 🛠 How writing things down—change docs, RFCs, and POCs—dramatically improves engineering quality 🧠 The mindset shift from coding features to shaping architecture and culture 🌿 How Erik maintains work-life balance while leading across continents and teams 🎧 Listen now: ▶️ YouTube: https://lnkd.in/dp3ftkMx 🎧 Spotify: https://lnkd.in/dcFFsGCR 🍏 Apple: https://lnkd.in/dG8Wmbwd 📬 Newsletter: https://lnkd.in/d3cUAzpJ #microfrontends #observability #softwarearchitecture #newrelic #softwareengineering #frontend #leadership #teammanagement #engineeringculture #señorsatscale

20

Why do I write here? For me, writing is a stress test for my ideas. When an idea lives only in my head, it feels brilliant. Clean. Fully formed. But the moment I try to put it on paper, the illusion breaks. I see the idea for what it actually is: incomplete, simple, or embarrassingly vague. Externalizing thinking exposes the gaps, and illuminates the solution path. I didn’t need a psychological study to learn this, I learned it by confronting my own thoughts on paper. My internal rule is simple: If I can’t explain it clearly enough to convince myself, I have zero chance of convincing you. But here’s the truth about my private writing: It’s a mess. If you saw my notes, you’d find scattered words, bilingual phrases, and arguments that only make sense if I’m standing next to you explaining the context. It’s just a raw dump of my brain, a serialization without "form". That is why I post on LinkedIn. Writing in public acts as a forcing function. It forces me to do two things I never do in private: 1. Clarify the idea until it stands entirely on its own. 2. Respect the reader’s time by focusing more on "form", not just content. Beyond that selfish motivation, there’s also a belief I hold: We all carry non-obvious, hard-earned experiences. Most of that knowledge stays invisible, and so many fascinating stories are lost because we never take the time to share them. If your brain works like mine, I encourage you to give writing on LinkedIn a shot. If even one person finds value in what you write, the effort is justified. And if not? At least you walk away with a clearer mind than you started with.

71

4 commentaren

I’ve been an engineer at WhatsApp for 6 years. In the past few weeks, I’ve gotten a lot of questions from friends in the Netherlands on how WhatsApp protects your privacy and I wanted to answer those here. Every message, photos, video, and call you make on WhatsApp is protected by end-to-end encryption using the industry leading Signal protocol, the same encryption technology used by Signal itself. Each message is encrypted using public keys, and can only be decrypted using a private key that resides on the recipient's device. WhatsApp does not have that key, which means that we cannot decipher your messages nor anyone else outside your conversation. We cannot decrypt user messages even if we wanted to do so. This is by design. There are no backdoors for anyone outside the conversation to access your messages. Simply put, your personal messages on WhatsApp stay between you and the people you choose to communicate with. For those who like to back up their messages to iCloud or Google One, we also launched end-to-end encrypted backups in 2021 so that your backups are protected too. Only you can access these backups, no one else, not us, Apple, or Google can do so. I’ve also heard questions about metadata, e.g. who you talk to, when, and for how long. Let me be clear: WhatsApp does not keep logs of who everyone is messaging or calling. We minimize the types of data that we collect and store, and only use the bare minimum to keep users safe from spammers and other harm.  I’ve been asked directly if we are tracking who people message for ads. We do not. Lastly, we have also taken a strong view against dangerous ideas proposed by some in Europe to introduce so- called “client side” or “device side” scanning of messages. Similarly, WhatsApp has spoken out against any attempts by governments to weaken security, often well intended in the name of child safety. We spoke against the UK government trying to force a backdoor on Apple because we believe it is wrong and dangerous. Protecting people’s personal messages is our most important job at WhatsApp and we’re entirely committed to this goal. WhatsApp is -and always will be- a place where you can have safe, private conversations, no matter where you are. You can learn more about our approach at https://lnkd.in/giiZS2cE

423

17 commentaren

Last night at the AI Meetup in Utrecht, we asked a simple but uncomfortable question: 👉 Why will AI adoptions stall at 2X instead of delivering the 10X impact everyone is hoping for? Here’s the truth we uncovered together: In a Resource Topology, AI accelerates silos — faster specs, faster tests, faster code, but outcomes stay stuck in queues and approvals. In a Delivery Topology, AI helps teams ship faster features, but every pivot requires another silo, another backlog, another painful reorg. Only in the Adaptive Topology does AI stop reinforcing dysfunctions and start amplifying learning, adaptability, and innovation. ⚠️ Faster tasks ≠ faster outcomes. ⚠️ Faster features ≠ adaptability. The evening ended with one clear insight: 👉 Your org design decides whether AI delivers 2X or 10X. Thanks to everyone who joined, and a shoutout to the organizers Thijs Assmann and Chiel van Ewijk for Capgemini and Agile Consortium.

34

3 commentaren