Editorial: More changes for comparisons

[jmdyck]

... in line with PR #2877 and PR #3202.

The first commit changes:
_X_ and _Y_ are [not] the same {Realm,Module} Record
to just:
_X_ is [not] _Y_

The second commit changes:
SameValue(_X_, _Y_) is {*true*,*false*}
to
_X_ is [not] _Y_
when the two values are both Strings, or else one is a String and the other is a String or *null*.

[michaelficarra]

this WFM though I think one of the other editors preferred to keep "the same X Record" when we talked about this last

[jmdyck]

... I think one of the other editors preferred to keep "the same X Record" when we talked about this last

It seemed odd to me that #2877 didn't eliminate that form, so I checked the PR's comments for any discussion of it, but didn't find anything.

[bakkot]

That sounds like the sort of opinion I'd have. I do kind of prefer "the same X Record" because it makes it really, really clear that comparison is by identity rather than structural, without you having to keep in mind the details of how identity works for each kind of spec value.

[jmdyck]

it makes it really, really clear that comparison is by identity rather than structural,

So are there other kinds of values for which you'd like this clarity?

[bakkot]

Only really Records and Lists.

[bakkot]

Per call, we mildly prefer to keep the "same Record as" form, extending I guess to using it in the existing Repeat, while _thisEnv_ is not _varEnv_ case.

[bakkot]

But we do like the change to use is instead of SameValue for comparing strings.

[jmdyck]

Okay, I've dropped the first commit (reinstating the status quo uses of are the same Foo Record), and I've added some more commits to align the spec with comparison conventions.

[jmdyck]

(force-pushed to squash it down to one commit)