- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Wed, 2 Dec 2015 08:54:49 -0800
- To: Domenic Denicola <d@domenic.me>
- Cc: Yves Lafon <ylafon@w3.org>, Boris Zbarsky <bzbarsky@mit.edu>, "public-script-coord@w3.org" <public-script-coord@w3.org>
On Wed, Dec 2, 2015 at 8:30 AM, Domenic Denicola <d@domenic.me> wrote: > From: Yves Lafon [mailto:ylafon@w3.org] >> Note that the goal is not to produce a document to be used for >> implementors of WebIDL (who should follow the edcopy), but a reference of >> what is implemented at the time of its publication for writers of other >> specifications (and the SoTD has a paragraph about this, it may be clearer or >> more visible) > > Given this as the goal, I object to the publication. We should not encourage other specifications to reference an unimplementable and unimplemented specification. And we should not publish such specifications in the first place. Strongly agree. The sole and only purpose of publishing an obsolete version of a spec is if the publication offers new and important patent protection, and even then we should try to publish the most up-to-date thing we can. Obsolete versions of specs are *not* appropriate for reference except as historical artifacts. They should not be referenced by living specs unless explicitly noted as a historical artifact that is not intended to offer implementation guidance, but merely to illustrate some point that is aided by historical evidence. This issue keeps coming up over and over again. What process changes do we need to make to ensure that obsolete versions of a spec are not published as a "current" version? ~TJ
Received on Wednesday, 2 December 2015 16:55:37 UTC