[jauntywundrkind]

The article feels so intolerant to complexity, desparages again and again the complexity and accumulation of concerns.

But amid that there is this wild sublime hopeful optimism. This is maybe the only even slightly pro-CORBA piece any person machine has written in probably 20 years now. It/they try to pitch WASI WebAssembly as the next coming, as some magical relief that may somehow perhaps (entirely unspecifiedly) cure the complexity issues of today up.

And WASI is cross language! You can make wasm modules in whatever language, throw them into your wit world, hit go, and get a runtime that's running all your languages.

> The problem with this thesis is that there’s no other easy way to make microservices composed of different language stacks work well together

This is actually far easier. Because who cares if you are running micro-services? Every microservice exposes some http services. With some schema (json schema, openapi, type spec, GraphQL, whatever). One of the real true gains of micro-services is that language is (mostly) irrelevant.

[lambdasquirrel]

I don’t think WASI is going to make k8s (and its ecosystem) go away. All the other problems tackled by docker and k8s were not solved by things like GraphQL and protobuf.

[jauntywundrkind]

Very much agreed. WASI feels like such a code concern, but there's so much more to manage. K8s is such an excellent extensible management layer, that I fully expect to endure & power Wasm.

Wasmcloud for example already does this very well.

Had some similar comments in my inordinately long post in this submission, https://news.ycombinator.com/item?id=44460641