jgraham

My proposal isn't a proposal as such, it's an observation about what would be most consistent going forward. That said, one reasonable question is how widely deployed SPC and Custom...

Camel case is spelled `camelCase`; `CamelCase` would actually be `PascalCase` (which would arguably be better than camelCase for emum varaiants since it's commonly used for type names and enum variant...

I think the camel case examples are mostly of things that aren't enum values. Capabilities in particular are just JSON keys, and all the event properties like `isComposing` aren't part...

I also think that at this point looking in BiDi is also helpful as precedent.

In WebDriver I see the following things that are actually enumish values: 1. camelCase * Action names 2. lower-case-dash-delimited * [none] 3. lower case space separated * Error codes *...

In WebDriver-bidi the list looks like the following (excluding things that direct ports from WebDriver): 1. camelCase * Method names (These are intended to look like JS methods, not enums,...

It occurs to me that one technical consideration is that `alllowercase` doesn't allow converting to other forms algorithmically, whereas all the others do allow that.

Arguably we should keep the issue open until we have a PR documenting the naming conventions in the spec.

I assume it's supposed to be https://w3c.github.io/pointerevents/#the-pointercancel-event which is when the pointer isn't going to produce new events e.g. because it was disconnected. I agree that there's currently a hole...

FWIW I'm happy with this only being returned, not used as the basis for a match.