yosh

let me ask you this — is dom diffing the bottleneck?

yeah, it could be - not usually but it's the extreme cases that would act violently; capping them without API change would be neat

in the super hot path i always go with vanilla JS and target the exact nodes i intend to update, but diffing/patching... 1/2

is in comparison, inherently inefficient (any way you look at it, its an abstraction), so why dom-diff when perf maters? 2/2

I respectfully disagree here - I think the logic to handle the custom updates may sometimes be heavier than a single diff per frame

you mean the complexity of the update may be so substantial that it justifies the abstraction?

in some cases: definitely - but this def depends a lot on who wrote the code for the update; feel like vdom may be more consistent

if updates are so complex it becomes a cognitive strain to apply - it's probably spaghetti 1/2