Re: [whatwg/webidl] Defining AbstractModuleSource intrinsic inheritance (Issue #1322)

My very personal preference is to define things systematically rather than defining things with ad-hoc rules (so, I tend to like the idea of [IntrinsicHeritage]). Having said that...

1. Once it's explicitly supported (showing the existence of [IntrinsicHeritage] explicitly), I'm afraid that people would start abusing it more and more. E.g. "Oh, we can use this feature for our own unique purpose! (although not in an originally designed way)". I feel it's hard to prevent people from using existing features even when the features are explicitly marked as obsolete, deprecated, unrecommended, unsafe, etc.
2. Given we have only two cases, I don't see the strong necessity of [IntrinsicHeritage].
3. I'm fine with a special requirement of HTMLAllCollection's [[Call]] being not supported by an extended attribute.
4. I'm fine with a special requirement of Location's "valueOf" property being not supported by an extended attribute.
5. ...

I understand I'm talking about a sort of ...politics?... and not design of software, architecture, etc. Purely technically talking, [IntrinsicHeritage] looks good. I understand %X.prototype% is different from %X%.prototype, but [IntrinsicPrototypeParent] looks a bit overkilling/redundant (we don't want to support totally different inheritances for interface objects and prototype objects but it looks possible). Can we implicitly assume %X.prototype% inheritance, too?

In summary, my preference is: 1. special rule (no extended attribute) is enough (and good in total imho), 2. either of [IntrinsicHeritage] or [IntrinsicParent] + [IntrinsicPrototypeParent].

-- 
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/webidl/issues/1322#issuecomment-1594252330
You are receiving this because you are subscribed to this thread.

Message ID: <whatwg/webidl/issues/1322/1594252330@github.com>