Post

Philip Jägenstedt
@foolip.org

This is the best explanation of the interplay of specs, tests, and implementations that I've read. The point of both specs and tests is interoperable implementations, but that doesn't mean that implementations are always right when they agree.

Brian Kardell@bkardell.com

My colleague @atbrakhi.bsky.social wrote a nice, cleverly titled post based on a very short exchange we had on Instagram the other day... www.atbrakhi.dev/blog/browser...

atbrakhi.devBrowsers Situationship: When Browsers Agree but the Spec Doesn’t | Atbrakhi's BlogWhat came first: browsers or specifications? Why are specifications important? Why do we sometimes change a specification instead of browsers? What matters most: the specification or the user experien...
Philip Jägenstedt
@foolip.org

I would take it one step further. I think that if two implementations agree on something, it's irresponsible to change it without first considering the shortest path to interop. "The spec says so" is not a good enough reason without that path.

0 reposts3 likes

Philip Jägenstedt
@foolip.org

In the end, often a combination of changes across specs, tests, and implementations are required. When collaboration is working well, it's a lot of fun to sort out obscure things like how newlines should be normalized, especially when web developers and users are affected by the lack of interop.