Hacker News new | past | comments | ask | show | jobs | submit login

First of all, congratulations on publishing your app! I’ve never come close to that, although having created a few PoC’s.

On my 11 Pro the animations are sluggish, buttons presses are missed when switching from plan to browse and the device gets pretty warm indicating heavy cpu usage. I’ve read that you’ve already put a lot of time in investigating performance issues, so is then expected with a more complex UI, is this the best one can do with SwiftUI, or do you need to be an even more experienced expert in SwiftUI to further improve things?




Thanks so much for the kind words!

Okay, so a few things:

1) we’re still in beta, and so we’re aware that the experience still has room for improvement in terms of optimisation. We’ve made big strides recently, but there’s still ways to go clearly. We won’t consider ourselves out of beta until we are happy with performance across the board. First wave of improvements imminent, and we’ll continue to between now and full release.

2) that said, based on our dogfooding, profiling and user feedback, your experience seems to be substantially worse than anything we’ve encountered/been made aware of. We have stress tested the app too with a calendar filled with an unusually high number of events on a daily basis, and although I wouldn’t describe performance as _good_, it’s borderline acceptable at worst. It would be great — if you are available — to learn more about your setup (bardi@timing.is) so we can have a better understanding of what could be tripping it up so badly for you. Also, as we make progress with optimisation work, it would be super helpful to use your device as a point of reference when performing more anecdotal before/after comparisons.

3) as suggested in the post, we wouldn’t describe ourselves as SwiftUI _experts_. So, I don’t for a moment want you to takeaway that this is the best SwiftUI can do when managing an app of moderate complexity. I’m confident we’ll be able to make inroads ourselves, but most definitely everything would be running smoother already if we had direct access to expertise. But again this highlights the main point that hopefully wasn’t lost: it shouldn’t require experts to make SwiftUI perform well, and — on the flip side — it shouldn’t be so easy to make SwiftUI perform poorly.

The other being, it shouldn’t be impossible to perform certain interactions that for a four version old product you would expect it to be able to support. I understand like any tech, you need to prioritise what it can do out of the gate, pushing forward most common use cases. At this point though, I shouldn’t need to jump through unclear hoops to get a scrollTo function to work in a LazyVStack.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: