Maciej Malinowski
91 posts



UWAGA - POTENCJALNA PRÓBA WYŁUDZENIA DANYCH PODCZAS WYBORÓW!!! Jeżeli ktokolwiek z głosujących z "Zaświadczeniem o prawie do głosowania" spotka się z próbą użycia tego typu nielegalnych aplikacji do sprawdzania czegokolwiek proszę od razu wzywać Policję @PolskaPolicja oraz powiadomić Państwową Komisję Wyborczą @PanstwKomWyb! Obwodowe Komisje Wyborcze mogą korzystać tylko z narzędzi zatwierdzonych przez PKW. Nikt nie ma najmniejszego prawa stosować nieautoryzowanych aplikacji które mogą być wyłudzaniem danych lub próbą zakłócania wyborów.




I have two spicy takes 🌶️ 1. Developer content is sometimes at odds with what's best for developers 2. Most "boilerplate" templates include way too much 1️⃣ The race to be first... There's an increasing trend of content being created the day a new developer tool or update is released. This race to first can be harmful. It's tricky. Developers are excited about new things. They want to stay up to date on the latest trends. But the *message* of the content creator can easily accidentally shift from "this is interesting and new" versus "the old thing is bad, this is better". In the rush to talk about something new, I see folks miss a lot of nuance in the discourse. Or maybe that new thing still has pieces to be figured out—that might be fine! I feel this is at odds with what's best for developers, because you should likely not immediately adopt the new thing. You might start to pay more attention to it, maybe read through the docs, but not put it in production right away. In reality, any non-trivial site/app needs an incremental path to the new thing, anyway. So what can content creators do? Present both the pros and the cons. Acknowledge the design tradeoffs. State when something is new and probably should not yet be adopted. Spend extra time before creating content to form a more rounded opinion. Build something small with the tool and present practical tradeoffs. 2️⃣ The template that never stops... A common starting point for newer developers are boilerplate/starter repositories that include a bunch of services/libraries hooked together in a single example. In my opinion, most of the time these are not helpful. This reminds me of the now-universally-painful ESLint configs that get forced on folks. All of these extremely opinionated changes were made, and I don't understand why. I frequently see starter templates that include entirely stylistic things to the point of diminishing returns. Do you really need conventional commits in your starter template for building a product? Probably not. Most importantly, you've just dropped dependency-hell on a new developer. These templates can be hard to maintain and often are a worse starting point that the minimum-viable-template. So what can we do? Smaller, more focused templates. If there's a pain point in getting two services/libraries to work together, that's a good starting point. It doesn't need 15 other things also included. That's all for today 😁






















