Revert "Rename SpanContext to SpanReference. (#1075)"

[carlosalberto]

Fixes #1126

I'd like to get the eyes of as many people as possible. We will also reserve a few days before (hopefully) merging this.

[jmacd]

This was discussed extensively in yesterday's Spec SIG (APAC) call. We found among the group general consensus that SpanContext is not a perfect term, but that the cost of changing from SpanContext (which is well established) is too high to justify the change.

[tigrannajaryan]

I don't have a strong opinion, approving to indicate I saw this and fine with the change.
I advise all @open-telemetry/specs-approvers and language maintainers to also indicate that they saw this even if you don't have any comments.

[obecny]

lgtm

[dyladan]

Seems fine to me. I think we've used the term "context" in far too many places. The cost of changing now is high, but what is the point of alpha/beta cycles if not to make expensive changes before GA?

[tsloughter]

Put my reasoning for approval in the issue #1126 (comment)

[anuraaga]

I still slightly favor SpanReference, but probably because it's more idiomatic in Java than other languages since everything's a reference anyways. For the record, this is the sort of code that's unfortunately bemuddled with two contexts, but I understand the motivation to not make a big change at this point

Span.fromContext(context).getSpanContext()

[Oberon00]

Approving because I agree the cost of renaming may be too high. I do think that SpanReference is the better name for that concept though.

[arminru]

Same as @Oberon00.
@carlosalberto you might want to indicate in the changelog that the change was reverted.

[carlosalberto]

@arminru Removed the original entry in the CHANGELOG regarding the rename to SpanReference ;) Hope that works!

[bogdandrutu]

Wow only 10 approvals :))

[carlosalberto]

Merging this as we have 1) enough approvals (7 approvers and 3 language SIG approvers), and 2) let it soak for a pair of days already.