Right, WPF's declarative databinding support is separated from imperative semantics, even if they can interact a bit (a databinding relationship can be formed and destroyed imperatively). But I've moved away from WPF to Glitch, which doesn't have that problem: all code runs in the same world, with dependency tracing, re-evaluation on dependency change, and logging to support rollback as needed. In that case, now I'm curious how React compares to what I'm working on right now.
I've never heard anyone claim that laziness leads to more performant UIs. It has always been the opposite in my experience: by loosening up evaluation orders and allowing for glitching, you reduce the amount of book keeping needed to come up with optimal DAG-style re-computation orders. For example, consider a double-indirect (flat-map-style) dependency:
countLabel.Content = folderView.Selected.Count;
If count changes, you update the label. But if selected changes, you have to stop observing the previous selected folder's count property and start observing a new selected folder's count property. Carefully tearing down and building back dependency chains is tough, but if you just play lose with the dependencies, it is actually quite easy to handle (you might get some spurious re-evaluations). Note that WPF can't handle this without lots of hackery, which is one reason I moved all my UI work to Glitch (the other is that WPF is totally inappropriate performance wise for writing compilers and language-aware editors). I'm curious how React handles indirect state? Or does it need to do dependency tracing at all (it wouldn't if its not incremental; I can't tell by looking at the website)?
I've never heard anyone claim that laziness leads to more performant UIs. It has always been the opposite in my experience: by loosening up evaluation orders and allowing for glitching, you reduce the amount of book keeping needed to come up with optimal DAG-style re-computation orders. For example, consider a double-indirect (flat-map-style) dependency:
If count changes, you update the label. But if selected changes, you have to stop observing the previous selected folder's count property and start observing a new selected folder's count property. Carefully tearing down and building back dependency chains is tough, but if you just play lose with the dependencies, it is actually quite easy to handle (you might get some spurious re-evaluations). Note that WPF can't handle this without lots of hackery, which is one reason I moved all my UI work to Glitch (the other is that WPF is totally inappropriate performance wise for writing compilers and language-aware editors). I'm curious how React handles indirect state? Or does it need to do dependency tracing at all (it wouldn't if its not incremental; I can't tell by looking at the website)?