A lot of the boldness stands though, even with the compatibility layer, and it's a "layer" more than a pivot for Deno. deno.json is still far simpler than package.json. Deno still takes a batteries included approach with smart defaults by default. Deno still pushes you toward a modern-standards "native" approach: ESM by default; ESM native libraries including a growing "standard library" on JSR; your dependency graph is still mostly an importmap you can also dump directly into a browser, too (even some of the compatibility shims with the npm ecosystem).
Deno still has a permissions model that is very different and far more opt-in than Node. This post makes a case for thinking of Deno's deep, native OpenTelemetry support as something very new and different from Node's approach, and clearly important to the future of application deployment and observability.
Technically Deno is still very interesting in technical promise, especially compared to Node, even with a larger compatibility layer.
You nailed it. Its incredible how many dunces here are lamenting how Deno abandoned everything, when all they really did was add a compatibility layer.
Maybe next time instead of name calling you could consider that other people simply have different perspectives to you?
I and others are lamenting that the compatibility layer removed incentive to help create a new JS ecosystem that isn’t layers of garbage piled on top of each other. That new ecosystem is what I wanted and Deno is no longer the path to it. If that makes me a “dunce”, so be it.
Even for Node projects I've started searching JSR first and comparing JSR scores of libraries. If a project advertises first-class JSR support now, I pay more attention.
The compatibility layer works in both directions here. JSR is a fascinating growing home of Typescript-first, typedocs everywhere, ESM-native ecosystem. Most of those install just as well into Node as Deno (though they are easier to use in Deno, by far). I want Typescript-first. I want ESM-native. I have a registry now that gives me some indicator if and how much a package author cares about those things.
The compatibility layer didn't remove the incentive for Deno (the Company) to help create a JS ecosystem that isn't just layers of garbage dependencies piled into trenchcoats, the compatibility layer encouraged Deno to expand their horizons and make Node better too by providing a Registry like JSR. I think JSR is a better way to promote a new, fresher JS ecosystem (for all) than restricting what sorts of packages are compatible with the Deno runtime. It's got gamified test scores! It's got provenance data and supply chain transparency!
If they hadn't added the compatibility layer to Deno too many people would have chicken-and-egg excuses that all the packages they need are in npm and never even tried Deno. With the compatibility layer in place Deno gets to shine in its own place of what it can do better than Node, then suggest you try JSR in all your hobby projects, then get you using JSR even in Production on Node packages until it is even easier to convince your day job to switch to Deno for better performance and increased security and deep OpenTelemetry support… Deno's still helping a lot to create a new JS ecosystem with fewer layers of garbage, compatibility or not with the old layers of garbage.
Deno still has a permissions model that is very different and far more opt-in than Node. This post makes a case for thinking of Deno's deep, native OpenTelemetry support as something very new and different from Node's approach, and clearly important to the future of application deployment and observability.
Technically Deno is still very interesting in technical promise, especially compared to Node, even with a larger compatibility layer.