16:09
<Michael Ficarra>
thoughts on removing the @@x notation in favour of %Symbol.x%?
16:37
<bakkot>
that's https://github.com/tc39/ecma262/pull/1314
16:37
<bakkot>
we were doing a bunch of other edits at the time and I wanted to not churn as much
16:37
<bakkot>
but things have been more stable for a while now, so I guess I am not as against it
16:38
<bakkot>
I would still mildly prefer not to but if you really want to push for it I won't stop you
16:50
<Michael Ficarra>
I'm with @littledan on this one: https://github.com/tc39/ecma262/pull/1314#issuecomment-622600422
16:50
<Michael Ficarra>
we should change it
17:05
<shu>
i prefer to move away from @@ notation for a very simple reason, which is too many chat systems, including GH, keeps pinging people when we type it
17:05
<shu>
so editorially moving away from that helps communication
17:06
<shu>
from a pure reading perspective i mildly prefer to leave it be
17:06
<shu>
but in practice we talk about symbols all the time as implementers and practitioners
17:11
<bakkot>
If we start hanging well-known symbols off of other built-ins it might get a little confusing
17:11
<bakkot>
but whatever I guess
17:11
<shu>
hm that's fair
17:18
<ljharb>
should i rebase and update that PR then?
17:18
<ljharb>
even if we had, like, Map.specialSymbol and stuff i'd still think %Map.specialSymbol% works fine
17:19
<ljharb>
i'm not sure i see the value in knowing at the reference site if an intrinsic is a well-known symbol or not - and if that's a concern then we could (and probably should regardless) make intrinsic refs link to their definition
17:29
<bakkot>
it's just that it's being used for property access
17:30
<bakkot>
99% of built-ins are functions and so DefineProperty(_O_, %BuiltIn%) looks pretty weird
17:31
<bakkot>
symbols are importantly different from other built-ins even though you can in principle refer to them in the same way
18:05
<bakkot>
https://github.com/tc39/ecma262/pull/3178