Developer-Apple App Store split
There's been a lot of talk around this recently, especially after the newly formed Developers Union stated it as the next item on their agena after free trials for paid apps, and like many who are trying to make their living from the app store I'd also like to see developer's share of the proceeds be increased somewhat from the 70% status quo. The move to 85% for subscriptions after the first year, introduced in 2016, was a good first step in this direction. I'm not sure what a fair figure is, but hopefully Apple sees fit to make a few changes to help devs, at least those at the lower end of the revenue scale.
Improved Swift Handling in Xcode
A common complaint is that Xcode's performance is considerably worse when writing a project in Swift than the equivalent is in Objective-C. It was true when Swift first launched and is still true today (although to be fair, it is leaps and bounds better than it was in those early days). I like Swift (not that I disliked Objective-C, gasp I know :P) and have finally embraced it as the language I'll be using for any new work, and I'd love to see Apple double their efforts in bringing their dev tools up to speed.
I don't know what the odds of this one are but now having spent a few weeks with an Apple Watch I've come to love the platform's complications - the tiny customisable views that sit alongside the time on the watchface - that provide very accessible, gancable information. While arguably less useful on the phone, I still think something similar would be very handy on iOS' lockscreen.
Live App Icons
This is an ancient request that will never happen, but wouldn't it be cool if, like in the complication idea above, apps (with user permission of course) could periodically update their icons to display useful information? Like the built-in Clock, Calendar and Weather apps do already. I understand why Apple hasn't opened up this Pandora's box as it has the potential to overwhelm the user, I think approached in the right way (a limited templating system perhaps?) it would be an awesome addition.
PIP on iPhone
The iPad has has picture-in-picture for years now, and while I can somewhat understand with screen real estate being more of a premium why it perhaps wasn't an obvious feature for the phone initially, the number of times I've wished it was there is ridiculous and I'd use the heck out of it if Apple could find a way.
When compared to what is offered by even remotely recent releases of Android, iOS' notification handling is a bit of a joke. Having barely changed since the platform's inception over 10 years ago, notifications come in and are presented to the user chronologically. Beyond being able to enable or disable at an app level there is very little in place to help triage, group or prioritise the vast number of notifications that come in on a daily basis. This needs to change. At this point I don't care whether Apple outright copies Android or does things differently but it's past time that more tools for notification management were available on the "most advanced mobile operating system", as Apple has often claimed. However, given they've been absent this long I'm not holding my breath.
Apple Watch Improvements
Having only very recently got myself an Apple Watch - with the novelty factor not yet worn off - I'm new to realising the potential power of this little platform. I say potential because while the hardware is excellent and very capable for something that can be strapped to one's wrist, the software in places lets it down. To be clear, the built-in in functionality is good and reliable (and reason enough for the watch to exist) but third-party apps are lacking, and seemingly fewer now than they once were. Having spent the last few weeks building my own little transit app (mainly for myself, but it may see the light of day at some point in the future) I've realised what a frustrating experience it can be, explaining the shortage of quality apps. The APIs are mostly well-thought through but are limited, not always reliable, and in parts poorly documented. The debugging process isn't any better, between things working differently on the simulator and the real device and the slow and flasky connection to physical hardware it can be rather painful. In short I'd like to see the good ideas in watchOS 4 tightened up with a focus on making existing frameworks rock-solid and the development experience as a whole a little more tolerable. Some choice new APIs would be nice also, to give devs just a little more power and freedom to make useful watch software.
Continued Development for macOS
This is low bar for the Mac, but it is already wonderful and very capable platform, and I honestly can't think of too much it needs. A new UI framework more in line with iOS' UIKit, as has been rumoured, would be nice. The addition of HomeKit would seem sensible, for platform consistency if nothing else. I love the Mac, and I hope Apple makes some solid under-the-hood improvements but really any sign of active development will probably be okay with me.
Mentioned for Completeness
The Apple TV is a good product, and while apps may once have been heralded as the future more content is really what the platform needs. It will be interested to see what, if anything, in regards to tvOS will mentioned in the keynote on Monday.