- From: Domenic Denicola <d@domenic.me>
- Date: Wed, 2 Dec 2015 18:01:00 +0000
- To: Yves Lafon <ylafon@w3.org>
- CC: Boris Zbarsky <bzbarsky@mit.edu>, "public-script-coord@w3.org" <public-script-coord@w3.org>
From: Yves Lafon [mailto:ylafon@w3.org] > Well, when the goal is to document what is currently implemented. That > seems to go against your "unimplementable and unimplemented” argument. What is currently implemented is changing daily. (Literally; I saw a Chrome patch land this morning.) Trying to nail that down in a "level 1" distinct from the current spec will only lead to outdated documentation. If you're interested in documenting what Chrome 47/Firefox 42/Edge 13 implements, I would suggest the best venue for that would be the little browser support boxes at the bottom of MDN pages, or maybe caniuse. Publishing a so-called "recommendation", meant for reference by other specifications, is counterproductive toward that goal and has harmful knock-on effects across the ecosystem. At least caniuse/etc. clearly document what browser versions they are trying to capture, instead of pretending to be a specification. > The ‘latest version’ is what is not implemented and can’t be as there are > unspecified things. There are bugs in the spec, that is true. All specs have them.
Received on Wednesday, 2 December 2015 18:01:30 UTC