Ask HN: What do you think about future of Android? | Hacker News Reader

Ask HN: What do you think about future of Android?

8 pointswsieroci7 comments

Hi,

Simple question: what do you think about future of Android? Any thoughts?

Best, Wiktor

Android needs two things:

1. Real compatibility instead of compatibility libraries. Compatibility libraries just serve to fragment the community and frustrate developers (try to use Facebook integration fragments in a 4.0+ application if you don't believe me).

2. C#.

Android needs to evolve away from Java. It needs a Swift - something richer, more expressive, focused on modern 64bit platforms and above all open.

That post literally makes no sense.

Java is neither 32 nor 64 bit, it is bitness agnostic to a large degree (or at least where it matters it is, datatypes aren't really a major advantage, and most of the new CPU registers can be utilised in the JVM without the code being aware or being recompiled).

You also say that they need something "richer" (whatever that means?), "more expressive," (huh?), and "open" (even though the JVM on Android ("Dalvik") is free and open source software).

Sorry but everything you said there is either confusing or wrong. You just wanted to bring up Swift and seem to know absolutely nothing about Java/JVC/Dalvik, except that it is "old" and therefore just assume it is "bad."

I actually think C# is a better language than Java (at least in their current iterations). However not for any of the reasons you cited, for real reasons that actually make sense.

Value types with operator overloading. Expressive and good for performance.

Open? Java is owned by Oracle.

C# is better and more expressive than Java. Especially the functional/declarative features added after the initial Java influenced release - async/await, var, linq etc.

I agree, it was pretty much buzzword fest. Here : I think the Android framework (Java + Android APIs + Android dev tools) should evolve towards something less verbose. Let's say I want to add a new type of selector in my ListView (for example for the active entry of the navigation drawer). I need to declare it in attrs.xml, use it in my layout and also to write custom versions of all the views that are going to use that selector in order to override onCreateDrawableState and inject it there, not to mention all the plumbing in different classes in order to get & set that state.. Ideally, I should be able to simply declare that state in my layout selectors and simply set it in the corresponding adapter.

There are many, many other examples where I need to modify many classes in order to do something relatively common. There are also issues with some java aspects like the enums that are syntactically very nice to use but not that good for your performances. I also feel like a caveman each time I have to write a bitfield. I should have a construct built into the language (or at least the IDE, I don't really care which one) that gives me the flexibility and type safety of enums while performing like a bitfield.

I don't know if getting rid of java is the answer here. I would not mind to, but it would be very traumatic for the Android ecosystem. I hope that these pain points will be fixed (I/O 2015 ? ^^). I don't think that Google will ever use C#. However nice it is, it is controlled by Microsoft, which is probably reason enough for Google to avoid it. There are other options though, Dart, Rust, Go, an entirely new language, ...

Android will own the low end mobile space. iOS will own the high end.

Either the mobile OS wars are over and Android and iOS won (I think the most likely scenario for the next 10 years), or there are major developments coming up that we (I) can't foresee that will change this dynamic. But for now, that's it.

Android and iOS have won the first world, it's done. Some contenders (Microsoft, Blackberry) try to grapple some market share, but unless Google or Apple do something very stupid it seems doomed to be a very hard & slow progression. Even with Microsoft throwing moneys at every major app makers, the windows marketshare make it very hard for windows phone development to be worth it.

If I had to make a gamble, I would say that Google is going to take over the remaining 5 billion people too. Apple and emerging markets don't go together at all. Microsoft & Firefox are targeting these markets but I feel that with Android One & the changes made in 4.4 in order to have acceptable performances on low end devices, Google has the formula to win. Nothing is done yet though.

Google has also decided to conquer other form factors. In order to do not frighten its partners, these products are called Android (and it is indeed the Android tech stack behind them). Google Wear is the first wearable device that makes sense to me. That does not mean that it will be a big hit, but it is a good start.

All these new hardware initiative (car, health, wearable, home, ...) from Apple & Google are very impressive. They have many ideas in common and both only work with mobile devices made by their company. There are probably at least some technical reasons behind that move but it is still very disappointing. I don't think that anybody in their right mind could be entirely satisfied with a car or home automation system that dictate which brand of smartphone you are going to be able to use.

As an Android dev, I am confident in Google's capacity to provide the tools I need. Even though I don't think that Google made ALL the right choices (lockscreen widgets still look like a very weird choice to me, they seem to be disappearing with L though), my most critical needs were fulfilled with the last i/o round of announcements with improved performances, first class design ('Holo' was really starting to look bleak compared to the other OSes) and the replacement of the terrible ListView by RecyclerView.

There are a couple of parts where Android is still weak from a dev point of view : -Even though Google is visually making many efforts in that area, the company does not have the tool making expertise that Apple or Microsoft have. Using IntelliJ was a very good move, the JetBrains guys know how to create an IDE. Right now we are in a weird phase where Google is slowly giving the finger to Eclipse and forcing us to migrate to Android Studio but it is still a bit immature. Staying with Eclipse means not benefiting from many new things, while migrating to AS means dealing with its bugs..

-Java is not as bad as its reputation can make you believe it is but there are real pain points (in the context of Android development at least, I don't really know the other sides of Java development to be honest). For example, Enums in java are syntactically awesome but for big codebases, they can contribute to bloating up your app if you use too much of them (each time you use an enum for the first time, all its members need to be instantiated and enum also generates some big methods). The alternative is to use bitfield flags or simple ints. It is better from a performance perspective, but it is absolutely ugly to use and does not offer any type safety. Some of Java weaknesses can be overcome with things like EventBus or RXJava but some are too deeply engrained in the language in order to be solved.

-Again, Google does not have enough tool-making expertise. I should be able to override the default widgets colors with a simple line in my layouts. This has just been added with the L release and I am not even sure that it is a back-compatible change, this should have been an API lvl 1 feature. There are many other examples where hundred of lines of code are necessary whereas it should be doable in one line with a good framework.

-Fragmentation is also not that bad but it is a reality you will have to work with. Crashlytics, bugsense and other crash reporting solutions are your best friends in order to detect and fix major fragmentation issues. Anecdotally, so far Samsung has been the biggest source of such issues for me.

-According to Android engineers, Java is here to stay. Since they don't communicate on their future plans, it is impossible to know whether they are working on alternatives or not but it would be welcome.

Reply on news.ycombinator.com