Embellish partial_implementation guideline#29661
Conversation
github-actions
Bot
added
docs
|
Tip: Review these changes grouped by change (recommended for most PRs), or grouped by feature (for large PRs). |
Elchi3
left a comment
Elchi3
left a comment
There was a problem hiding this comment.
Thanks Daniel, this is very helpful.
caugner
left a comment
caugner
left a comment
There was a problem hiding this comment.
LGTM, except for the bit about "Many bug report[s] filed by web developers".
| Strong signals of negative impact include: | ||
|
|
||
| - Bug reports filed by web developers shortly after a new implementation has shipped | ||
| - Many bug report filed by web developers or many comments from web developers on existing bug reports |
There was a problem hiding this comment.
Found a typo: 🥇
| - Many bug report filed by web developers or many comments from web developers on existing bug reports | |
| - Many bug reports filed by web developers or many comments from web developers on existing bug reports |
But most importantly, I'm not sure about:
- "Many" -> this suggests a single bug report isn't sufficient?
- "filed by web developers" -> this suggests bugs filed by the engine developers themselves don't count?
How about re-framing this as:
| - Many bug report filed by web developers or many comments from web developers on existing bug reports | |
| - Bug reports filed later with significant attention from web developers (through upvotes, comments, blog posts, or social media posts) |
There was a problem hiding this comment.
(Maybe "demonstrable attention"?)
There was a problem hiding this comment.
What I'm trying to get at here is that multiple developers have confirmed that the bug impacts them. They can do that by:
- Filing multiple independent or closely related bugs
- Commenting or upvoting on one existing bug
- Complaining in some other linkable venue (e.g., social media)
And yes, engine developers filing bugs is not a (strong) signal of impact on developers, absent other information (e.g., "I, engine developer, filed this bug because of a complaint I saw on social media").
There was a problem hiding this comment.
My main point is that I would expect there rarely to be "many" reports for any bug, there's usually one and maybe a few duplicates, but not many. Maybe the focus could be on evidence that this affects many web developers.
There was a problem hiding this comment.
OK, I've revised the text to avoid the word "many." See 3da308f.
Co-authored-by: Claas Augner <495429+caugner@users.noreply.github.com>
|
@ddbeck I'll let you merge in case you want to revise more. |
4 participants
Summary
This adds some more detail to our partial implementation guideline, covering second opinions and negative impact on developers.
Test results and supporting details
Based on a discussion on the BCD call today, about how to improve our process on setting and reviewing the
partial_implementationflag.Related issues
AuthenticatorAttestationResponse.getPublicKey()#29458getPublicKey(): add per-algorithm subfeatures #29637webauthn-public-key-easyto regress to limited availability web-platform-dx/web-features#3995