Re: [whatwg/dom] Proposal: DOMChangeList (#270)

It seems like there's general agreement that some kind of transactionality is desirable, but with a key unresolved question being whether we should strive to avoid exposing intermediate states. I find the idea of exposing those states pretty concerning, and I think it's worth trying to achieve an API that avoids it, which AIUI is what this proposal is aiming at.

I also find the idea of transactional change records an appealing primitive for modeling the DOM -- it seems like a good "explain the platform" approach: an execution trace is a sequence of DOM changes. So where other APIs @esprehn mentions might use wrappers to talk about existing nodes, it makes sense for transactional change records to operate at a lower conceptual layer.

Also, I'd be pretty worried about arguments saying that absent empirical data, we should have our primitives do more things and bet on our optimizations getting smarter. I think our default orientation should be the reverse: our primitives should be simpler and try to do less.

---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/dom/issues/270#issuecomment-227261146