Mobile Application Development

Mobile development life cycle

The mobile application development life cycle is the sequence of phases for building a mobile app.

1. Ideation and requirement analysis. Identify the problem, target users, platforms (Android, iOS), device constraints, features, competitors and business goals.
2. Planning. Select the technology (native, hybrid, cross-platform), team, schedule, cost and risk, and create the product roadmap.
3. Design (UI/UX). Create wireframes, mock-ups and prototypes, define navigation and the design system, and consider screen sizes, touch interaction and usability.
4. Architecture and development. Design the architecture (front end, back end, APIs, database), then code the app in iterations. Integrate services such as GPS, camera, push notification, payment and cloud.
5. Testing. Unit, integration, functional, usability, performance, security, compatibility testing on emulators and real devices, and beta testing.
6. Deployment (release). Publish to Google Play or the App Store after meeting their guidelines, with store listing, screenshots and versioning.
7. Maintenance and updates. Monitor crashes and reviews, fix bugs, add features, support new OS versions and analyse usage.

Mobile computing architecture

Mobile computing is usually described by a three-tier architecture.

1. Presentation tier (client). The user interface on the mobile device (native app, mobile browser or middleware client) which handles input and display.
2. Application (middleware) tier. Business logic, application servers, web services/APIs, authentication, data transformation and gateways (for example WAP gateway), which also hides the differences between devices and networks.
3. Data tier. Databases and back-end systems (cloud storage, enterprise systems), with local storage on the device for offline use and synchronisation.

Wireless networks (cellular, Wi-Fi, Bluetooth) connect the tiers, and the architecture must handle limited bandwidth, intermittent connectivity, battery, security and mobility.

Elements of user interface

A mobile user interface is built from the following elements.

1. Input controls. Buttons, text fields, checkboxes, radio buttons, drop-down lists, toggles, sliders, date pickers and search boxes.
2. Navigation components. Tab bars, navigation drawers (hamburger menus), breadcrumbs, back buttons, bottom navigation and gestures such as swipe.
3. Informational components. Labels, icons, notifications, progress bars, toasts, dialogs, badges and tooltips.
4. Containers and layout. Lists, cards, grids, scroll views and screens that organise content.
5. Visual elements. Colour, typography, images, spacing, and animation.
6. Content. Text, images, audio and video.
7. Interaction. Touch, gesture, voice and haptic feedback.

How to design the right UI

• Know the users. Research their needs, context of use, and abilities, and build personas.
• Keep it simple and consistent. Use a clear hierarchy, consistent colours, fonts, icons and patterns; follow platform guidelines (Material Design for Android and Human Interface Guidelines for iOS).
• Design for small screens and touch. Touch targets of at least about 44-48 px, thumb-friendly layout, the important content at the top, and minimal typing.
• Easy navigation. Few taps to reach goals, a clear back route and visible current location.
• Give feedback. Show loading, success and error states, with helpful messages.
• Be responsive. Adapt to different screen sizes, orientations and densities.
• Accessibility. Sufficient contrast, scalable text, labels for screen readers.
• Performance and context. Quick loading, offline support, battery and network efficiency.
• Prototype and test. Create wireframes and prototypes, run usability tests, collect feedback and iterate.

Mobile TCP

Standard TCP assumes that packet loss is caused by network congestion, so it reduces its sending rate. In wireless networks, loss is often due to bit errors, fading and handoffs, so TCP performs poorly. Mobile TCP variants improve this.

• Indirect TCP (I-TCP). The connection is split into one over the fixed network and another over the wireless link at the base station or foreign agent. Wireless problems are hidden from the fixed sender, but end-to-end semantics are lost.
• Snooping TCP. A base station agent buffers packets and locally retransmits lost ones, and suppresses duplicate ACKs, keeping end-to-end semantics.
• Mobile TCP (M-TCP). Splits the connection using a supervisory host and a wireless side; when the mobile is disconnected, it freezes the sender's window (zero window) rather than triggering congestion control, and resumes at full speed on reconnection.
• Other methods: fast retransmit/fast recovery after handover, selective acknowledgements, and transaction-oriented TCP.

Methods for testing mobile applications

1. Functional testing. Checks features, workflows and business logic.
2. Usability testing. Evaluates ease of use, navigation and screen layout with real users.
3. Compatibility testing. Tests on different devices, screen sizes, OS versions and browsers.
4. Performance testing. Load, stress and response time, memory and battery consumption.
5. Security testing. Data storage, authentication, encryption and vulnerability checks.
6. Interrupt testing. Calls, SMS, notifications, low battery and network switching.
7. Network/connectivity testing. 2G/3G/4G/Wi-Fi, offline behaviour and handover.
8. Installation and update testing. Install, upgrade, uninstall.
9. Localisation testing for language and region.
10. Methods and tools: emulators/simulators, real devices and cloud device farms, manual and automated testing (Appium, Espresso, XCUITest) and beta testing.

Frequency spectrum

The frequency spectrum is the range of electromagnetic frequencies used for communication, divided into bands which are allocated by regulators (ITU, and in Nepal the Nepal Telecommunications Authority) to services such as radio, TV and mobile telephony. It is a limited resource, and mobile operators use licensed bands such as 900 MHz, 1800 MHz and 2100 MHz.

Comparison of 2G, 3G and 4G

Maps and location features are integrated in a mobile app using location providers and mapping APIs.

Steps (Android as an example)

1. Obtain an API key. Register the project with Google Cloud and enable Maps SDK and Places API (or use OpenStreetMap/Mapbox).
2. Add the library and permissions. Add the Google Play services Maps/Location dependency, and request ACCESS_FINE_LOCATION or ACCESS_COARSE_LOCATION at runtime with the user's consent.
3. Display the map. Add a MapView/SupportMapFragment to the layout and obtain the GoogleMap object in onMapReady().
4. Get the user's location. Use the Fused Location Provider (GPS, Wi-Fi and cell data), request location updates or the last known location.
5. Add features. Markers and info windows, camera movement and zoom, polylines for routes using the Directions API, geocoding and reverse geocoding (address to coordinates and vice versa), place search, and geofencing.
6. Handle practical issues. Battery use (reduce update frequency), accuracy, offline maps, permissions denied, and privacy.

On iOS, MapKit and Core Location serve the same purpose. Cross-platform frameworks provide equivalent plugins.

Mobile application development frameworks provide tools, libraries and APIs for building apps. Three examples:

1. Android SDK (native Android). Official framework from Google using Kotlin or Java with Android Studio. It provides full access to device features and best performance, with UI built by XML or Jetpack Compose, but the app runs only on Android.
2. Flutter. Google's open source cross-platform UI toolkit using the Dart language. A single codebase compiles to native code for Android, iOS, web and desktop, with a rich widget library, hot reload and near-native performance.
3. React Native. Facebook (Meta)'s framework using JavaScript/TypeScript and React. A single codebase produces apps for Android and iOS that use native UI components, with a large community, hot reloading, and reusable components.

Other examples are iOS (Swift and Xcode), Xamarin/.NET MAUI, Ionic (hybrid, web technologies with Cordova/Capacitor) and Apache Cordova.

Synchronization

Synchronization keeps data on the mobile device consistent with data on the server (or other devices). Because mobile devices are often disconnected, the app works on a local copy while offline and then synchronises when connected. Steps: detect changes (timestamps, version numbers or change logs), exchange changes, resolve conflicts (last writer wins, server wins, merge or user decision), and update both sides. Approaches include one-way and two-way sync, full and incremental sync, and manual, scheduled or push-triggered sync (for example SyncML).

Replication

Replication means keeping copies (replicas) of data on multiple nodes, such as the central database and the mobile devices, so that data is available locally even when disconnected, and performance and reliability improve.

• Types: master-slave (one-way) and multi-master (peer-to-peer) replication; synchronous and asynchronous; full or partial (only the data needed by the user).
• Issues: consistency, update conflicts, bandwidth and battery use, security and storage limits.

Together, replication provides the local copies and synchronisation keeps them consistent, giving mobile users seamless access to up-to-date data.

Generic UI

A generic UI is a device-independent user interface description that is written once, in an abstract form (using standard components such as text, list, button and form) and then rendered appropriately for different devices, screen sizes and platforms by a transformation layer (for example XML-based UI description languages). It avoids rewriting the interface for every device, and gives consistency, but may not use the special features of each device.

Advantages of voice UI on mobile apps

• Hands-free and eyes-free use. Useful while driving, cooking or walking.
• Faster than typing on small keyboards, and quick commands.
• Accessibility. Helps visually impaired users and those with difficulty typing, and supports low-literacy users and local languages.
• Natural and convenient interaction. Conversation-like and needs little learning.
• Better engagement and productivity, for example voice search, dictation and assistants.
• Works with small screens or no screen (wearables, smart speakers).

Limitations include noise, accents, privacy and accuracy.

Non-GPS solutions

Non-GPS location solutions determine the position of a mobile device without a satellite receiver, which is useful indoors and where GPS signals are weak or the battery is limited.

• Cell ID / cell-of-origin. The position of the serving base station.
• Triangulation / trilateration. Using signal strength, Time of Arrival (TOA), Time Difference of Arrival (TDOA) and Angle of Arrival (AOA) from several base stations.
• Enhanced Observed Time Difference (E-OTD) for GSM.
• Wi-Fi positioning. Using known access point locations and signal strength (fingerprinting).
• Bluetooth beacons and RFID for indoor positioning.
• IP address and sensors (accelerometer, gyroscope, compass) for dead reckoning.

Accuracy is generally lower than GPS (from tens of metres to some hundreds of metres).

Mobile ecosystem

The mobile ecosystem is the set of interdependent players that deliver mobile products and services:

• Device manufacturers (Samsung, Apple, Xiaomi).
• Operating system and platform providers (Android, iOS).
• Network operators (Nepal Telecom, Ncell).
• App developers and app stores (Google Play, App Store).
• Content and service providers, advertisers, payment providers.
• Infrastructure and chip vendors, regulators and standards bodies.
• Users, who create demand.

Their cooperation produces value for all participants.

Wireless Application Protocol (WAP) enables mobile devices to access internet content over low-bandwidth wireless networks. Since the mobile browser cannot interact directly with ordinary web servers, intermediaries are used.

WAP gateway

The WAP gateway sits between the wireless network and the internet (web/origin server). It:

• Translates between the WAP protocol stack (WSP, WTP, WTLS, WDP) and the internet stack (HTTP, TCP/IP).
• Encodes content: converts WML into compact binary form for efficient transmission, and compresses data.
• Handles requests from the mobile client, fetches the content from the web server and returns the response.
• Provides security (WTLS to SSL/TLS translation), user identification, and caching.

WAP proxy

A WAP proxy acts on behalf of the mobile client, accepting WAP requests, forwarding them to origin servers and returning the responses. It combines the protocol gateway and content encoder/decoder functions, and may also provide caching, content filtering, and customisation for the device.

Flow: Mobile device (WAP browser) sends request over the wireless network to the WAP gateway/proxy, which converts it to HTTP for the web server; the server's HTML/WML response is processed by the gateway into compact binary WML and delivered to the device.

Form factors

The form factor is the physical size, shape and layout of a device and its input/output design. Mobile form factors include:

• Bar (candy bar) phones with a fixed screen and keypad.
• Flip (clamshell) and slider phones.
• Smartphones with a large touchscreen.
• Phablets and tablets with larger screens.
• Foldable devices.
• Wearables such as smartwatches and fitness bands.
• PDAs and handheld/QWERTY keyboard devices.

They affect screen size, resolution, input method (keypad, touch, stylus, voice), battery, and so the UI design and usability.

Text to speech (TTS) techniques

TTS converts written text into spoken output. The stages are text analysis and normalisation (expanding numbers and abbreviations), linguistic analysis (phonetic transcription and prosody), and waveform generation. Main techniques:

1. Concatenative synthesis. Joins recorded speech units (phonemes, diphones or words) from a database; natural-sounding but large in storage (unit selection).
2. Formant synthesis. A rule-based synthesis of speech using acoustic models (formants), without recordings; small and flexible but robotic.
3. Articulatory synthesis. Models the human vocal tract.
4. Statistical parametric synthesis (HMM-based). Generates speech parameters from trained models.
5. Neural (deep learning) TTS. For example WaveNet and Tacotron, producing very natural speech.

TTS is used for screen readers, navigation, assistants and reading messages aloud.

Mobile IP network

Mobile IP is the IETF standard protocol (RFC 5944) that lets a mobile node keep the same IP address while moving between networks, so that ongoing connections continue.

• Home address and home agent (HA). The node has a permanent address in its home network; the HA is a router there that intercepts packets for it.
• Foreign agent (FA). A router in the visited network that gives a care-of address (COA).
• Registration. The mobile node registers its COA with the HA.
• Tunnelling. The HA encapsulates packets addressed to the home address and sends them to the COA; the FA (or node) decapsulates and delivers them. Replies can go directly to the correspondent node.

GSM vs CDMA