Re: [dom] MutationObserver with querySelector for elements (#77)

I don't think such an investigation is necessary to justify the API. A high noise-to-signal ratio itself is a problem worth solving, and we've always wanted to add a filtering mechanism for MutationObserver from the beginning.  We never got around to it because each of us were distracted by other things.

In addition, without doing any investigation, I can tell you creating a JS array with thousands of entries is expensive and will cause a GC churn.  The fact MutationObserver records are objects is already bad, and we spent a lot of time minimizing the amount of information being collected for performance.  This is why we have options to collect old values for CharacterData and attribute value.  There is literally nothing controversial / mysterious here about the cost.

---
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/dom/issues/77#issuecomment-147483842