Accelerate your key learnings of scaling modern Android apps

Building Modern Android apps , best practices @Scale

Transcript

1. Building next generation apps - all-in one Accelerate your key learnings of modern Android app development with KMP Adit Lal GDE Android
2. Android GDE 🛠 Architect | Entrepreneur 🔊 Speaker 🌎 Globe Tro tt er 🍻 Beer enthusiast 🎯https:/ /androiddev.social/@aditlal 🔗aditlal.dev Adit GDE
3. Some early App development advice. Feature, Tech debt, and tips. How will Jetpack compose be a strong investment. Should you think in Kotlin Multiplatform. Our Talking points …
4. • Set clear goals for your Android app • De fi ne your target audience • Monetization & ROI Thoughts on App development
5. • Find out users' preferences and aversions. • Invest on a well thought process. • Focus on micro transactions - the core UX that gets your user to keep returning back to the app. Thoughts on Process
6. app core feature 2 feature 1 Monolith Expensive Object
7. What could possibly go wrong?
8. app core Feature B Feature A Feature modules Expensive Object
9. • Organise and decouple independent pieces of functionality • Improve project build times • De fi ne clear ownership boundaries between di ff erent teams Thoughts on modularisation
10. Feature module : • Owned by single team ? Set of team members • Encapsulates single feature • Single entry point • Smaller is better • Cannot depend on other feature modules • Every feature has an implementation and a contract module Thoughts on modularisation
11. App module : • No feature-speci fi c code • No infrastructure-speci fi c code. • Creates the DI tree structure and init it Thoughts on modularisation
12. Ownership? core auth common Feature n-1 Feature n+2 Feature n-2 Feature z Feature n Feature n+1 app … … … … …
13. Ownership? core auth common Feature n-1 Feature n+2 Feature n-2 Feature z Feature n Feature n+1 app … … … … …
15. Inversion of control Dependency Inversion Dependency Injection IoC Principle
16. • Declare Components that can be reused. • High level modules should not depend on low level modules • Build abstractions so any implementation can be easily changed with new one. • Dependencies can be injected into components. IoC Principle
17. Feature development
18. Serve webp images Make design easier to use. Think it through, vet libraries, and keep pe rf ormance in mind. Prioritise bandwidth of user network. O ff l ine fi rst is really powe rf ul UX experience.
19. 19 Build Speed Big Tech debt a ff ecting almost everyone
20. 20 Tech Debt Solve this at the most convenient time possible and o ft en. Don’t stack, solve it.
21. 21 Reduce background tasks when the device is running on ba tt ery power. Avoid wake-locks Batch calls - Work manager to rescue Don’t be greedy with GPS usage be proactive Minimize the sensors to when most needed Ba tt ery Historian - a great tool to benchmark app for ba tt ery usage Consider your footprint
22. 22 Jetpack compose
23. 23 Jetpack compose UI Toolkit is independent to the OS 🪄 State Management is built in🥳 Lots of context switching 🎉 So rt ed code logic 🚀
24. Android GDE 🛠 Architect | Entrepreneur 🔊 Speaer 🌎 Globe Tro tt
25. a wild idea to build apps pops! develop for Android and iOS
26. ‣ hire a team - Android and iOS ‣ de fi ne requirements ‣ plan roadmap for development ‣ develop and maintain and test apps ‣ production validation ‣ cross-check parity among both pla tf orms
27. native ~2x the team the cost time spent in meetings time spent on development time spent on testing time spent on bug fi xing
29. how do we have the best of both worlds?
30. a wild idea to build apps pops! and web develop for Android and iOS
31. native + web ~3x the team the cost time spent in meetings time spent on development time spent on testing time spent on bug fi xing
34. let’s fi nd a solution that works on all pla tf orms.
35. let’s fi nd a solution that works on all pla tf orms.
36. 2006 2008 2011 2012 2013 2015 2017 2018 . . . . . . .
37. Cross-pla tf orm advantages
38. Cross-pla tf orm advantages • smaller team/s • typically half of the cost needed on native • small learning curve for web developers
39. Cross-pla tf orm advantages dis • Locked to the framework implementation of UI - new updates from the OS will take time to adapt • pe rf ormance is not on par to native • One may end up also writing native code alongside • OS / device features are dependent on the fw suppo r
40. how do we have the best of both worlds?
41. how do we have the best of both worlds?
42. Kotlin • developed by JetBrains • open source • concise, safe, interoperable, dev tool-friendly • Suppo rt ed/used by Google for Android Development • It is so much more than “just for Android”
43. Kotlin K otlin multipla tf orm Using kotlin in projects to target more than one pla tf orm
44. Kotlin multipla tf orm “Shared UI is a history of pain and failure. Shared logic is the history of computers.” - Kevin Galligan
45. Kotlin multipla tf orm - language features - kotlin fi rst! - low risk - you decide what’s wo rt h to share across projects - interoperability - consistency across pla tf orms - strong community suppo rt
46. Kotlin multipla tf orm /kot · lin mul · ti · plat · form / / kät' lin malti' platfôrm, kät'lin melti'platfôrm/ noun optional, natively-integrated, open-source, code sharing pla tf orm, based on the popular, modern language kotlin. facilitates non-ui logic availability on many pla tf orms.
47. Kotlin multipla tf orm
48. Kotlin multipla tf orm
49. Kotlin multipla tf orm
50. Kotlin multipla tf orm but how can they expect to communicate?
51. Kotlin multipla tf orm but how can they expect to communicate? expect declared at common module actual declared at android module declared at iOS module declared at ...
52. Kotlin multipla tf orm Platform specif i c names?
53. Kotlin multipla tf orm Pla tf orm | common
54. Kotlin multipla tf orm commonMain src/commonMain/sample/Platform.kt expect object Platform { val name: String }
55. Kotlin multipla tf orm src/commonMain/sample/Platform.kt expect object Platform { val name: String }
56. Kotlin multipla tf orm src/commonMain/sample/Platform.kt expect object Platform { val name: String } targets: android and iOS or macOS de fi ne actual on android de fi ne actual on iOS de fi ne actual on macOS
57. Kotlin multipla tf orm src/androidMain/sample/Platform.kt actual object Platform { actual val name: String = "Android" } pla tf orm-dependent code
58. Kotlin multipla tf orm pla tf orm-dependent code actual object Platform { actual val name: String = “iOS" } src/iOSMain/sample/Platform.kt
59. Kotlin multipla tf orm src/androidMain/sample/Platform.kt actual object Platform { actual val name: String = "Android" } pla tf orm-dependent code actual object Platform { actual val name: String = “iOS" } src/iOSMain/sample/Platform.kt
60. Kotlin multipla tf orm pla tf orm-dependent code actual object Platform { actual val name: String = “macOS” } src/macosMain/sample/Platform.kt
61. Kotlin multipla tf orm pla tf orm-dependent code import io.ktor.* fun main() { embeddedServer(Netty, 9090) { routing { get("/hello") { call.respondText("Hello, API!") } } }.start(wait = true) } src/jvmMain/kotlin/Server.kt
62. Kotlin multipla tf orm
63. Kotlin multipla tf orm Actual needs to match because obviously it does
64. Kotlin multipla tf orm
65. Kotlin multipla tf orm
66. Common Kotlin Kotlin/JS JS code Kotlin/JVM JVM code Kotlin/Native Native code Kotlin multipla tf orm
67. Kotlin multipla tf orm Kotlin How much do we share Multipla tf orm What makes most sense
68. Kotlin multipla tf orm - strong community - a lot of people are using kotlin nowadays - Google and JetBrains are pushing Kotlin exclusively - you can ask questions directly at h tt ps://kotlinlang.slack.com conclusions
69. Kotlin multipla tf orm - 2x faster to develop your features business logic - 2x faster writing unit tests - One tech stack - consistency achieved across pla tf orms conclusions
70. Kotlin multipla tf orm - sta rt small, don’t try to reach 100% of shared logic - be extremely careful with all updates - IDEA, JDK, kotlin, libraries - everything. - keep versioning in mind Suggestions
71. Kotlin multipla tf orm Kotlin Slack Kotlin O ffi cial Kotlin Training Kotlin by: - h tt ps://jakewha rt on.com/presentations/ - h tt p://antonioleiva.com/kotlin Resources
72. Kotlin multipla tf orm h tt ps://github.com/joreilly/PeopleInSpace About Kotlin Multipla tf orm project with Swi ft UI, Jetpack Compose, Wear Compose, Compose for Desktop, Compose for Web and Kotlin/JS + React clients along with Ktor backend. Resources
73. Thats all folks! https:/ /linktr.ee/aldefy