Joshua MacDonald

I would be comfortable with a specification saying that everything is encoded in utf8, and that when truncating an SDK is permitted to use byte-level truncation, leaving an invalid code...

I second @reyang's opinion, speaking as the former maintainer of a multi-language stacktrace parsing library ([part of Google's offering in this space](https://cloud.google.com/error-reporting)) The first time we discussed adding a stacktrace...

I expected we would maintain two histogram implementations because we have two point kinds, and neither can exactly implement the other. However, I'd be glad to let explicit-boundary histograms be...

@tsloughter I wonder if there is a way you could adopt an existing stacktrace format like NodeJS or Java has?

I support both `system.uptime` and `process.uptime` semantic conventions.

I endorse @jack-berg's interpretation https://github.com/open-telemetry/opentelemetry-specification/issues/2810#issuecomment-1251651762. However, we must address the question posed by @MrAlias regarding #2619 concerning the precedent set in #2404. The key paragraph appears just above the edits...

This discussion was postponed until the 10/4 Spec SIG meeting.

This is partly covered in #2854. I believe we can close this issue after a suitable issue describing the confusion over exporter-related environment variables.

Another reason for this semantic convention worth mentioning: There is an interest in translating Prometheus Remote Write streams into OTLP streams, where data points with Cumulative start time SHOULD have...

Yes. I agree that both specifications are good to have. `process.uptime`: defined as a non-monotonic counter to signal that reset is not meaningfully permitted `process.start_time`: an attribute with a start...