Re: [w3ctag/design-reviews] JS Self-Profiling API (#366)

> That's an interesting but wondering if it might result in some unexpected leaks around CPU speed, etc. I'm happy you already mention timing. Let's look in more detail?

Absolutely. We don't anticipate this providing more timing information than `SharedArrayBuffer`/`performance.now()` ([see slides here regarding a polyfill](https://github.com/WICG/js-self-profiling/blob/master/doc/tpac-2018-slides.pdf)), but UA-specific mitigations may be necessary, as with any timing source.

> Another question, what's the intended use case. Do you foresee it used against live users, like in A/B testing (without the user being aware), or perhaps in a more testing-oriented environment in developer settings? Did you consider using permissions?

The intended use case is for live users in the wild, to get representative traces. Permissions aren't a great fit for this use case, particularly because that biases sampling (limiting the ability to collect profiles from the first load, as well as selecting for users with a higher propensity to accept permissions).

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/w3ctag/design-reviews/issues/366#issuecomment-496668592