Measurable CSSDB Stages

[jonathantneal]

How would we feel about agreeing to a measurable way to qualify the "stage" of a feature in the CSSDB?

Stage 0: Aspirational

I think this stage is well defined. Here is how I would qualify it:

• The feature is defined in a Unofficial Draft or Editor’s Draft specification.
• The specification exists on a public URL.
• The specification authorship belong to a W3C Working Group Member.
• The source code for the draft can be traced to a commit by that W3C Working Group Member.

Stage 1: Experimental

I think this stage is well defined. Here is how I would qualify it:

• The feature is defined in a Editor’s Draft or Working Draft specification.
• The specification exists on a public URL whose domain is csswg.org or w3.org.

Stage 2: Allowable

I think this stage is not well defined. Here is how I would qualify it so that it is well defined:

• The feature is defined in a Editor’s Draft or Working Draft specification.
• The specification exists on a public URL whose domain is csswg.org or w3.org.
• A previous version of the specification (from the Previous Versions metadata) exists on a public URL whose domain is w3.org.
• The feature is defined in the previous version of the specification.

An example of this is Custom Media Queries, which are listed as Stage 1, but would actually be Stage 2. ^reference

Stage 3: Embraced

I think this stage is well defined. Here is how I would qualify it:

• The feature is defined in a Candidate Recommendation specification.
• The specification exists on a public URL whose domain is w3.org.
• The feature is implemented or intended to be implemented by at least 2 recognized browser vendors, where implementation is known from MDN data, or where "intent to implement" is communicated in either a recognized "Platform Status" domain or a public GitHub Issue/PR URL (if someone really catches it fast).

Stage 4: Standardized

I think this stage is well defined. Here is how I would qualify it:

• The feature is defined in a Recommendation specification.
• The specification exists on a public URL whose domain is w3.org.
• The feature is implemented by all recognized browser vendors, where implementation is known from MDN data.

Rejected

I think this non-stage is well defined, and I would qualify it with the exact wording it already has:

• Any specification that has been rejected or neglected by its editor, or formally rejected by a W3C Working Group.

[Antonio-Laguna]

@jonathantneal this is great and gave me great insight. I've created a PR for CSSDB here: csstools/cssdb#78

[romainmenke]

We now also have minimumVendorImplementations in postcss-preset-env.

This is measurable and updates automatically based on data from mdn/caniuse.

---

We haven't updated the stage data in a long time 🤔