Supporting track

Mobile Automation (Appium)

One framework, two platforms, stable locators.

What this track covers

Appium architectureDesired capabilitiesUiAutomator2 & XCUITestAccessibility idsGesturesHybrid app contextsReal device vs emulatorParallel mobile execution

Cross-platform page objects

Keep one page object with platform-specific locator annotations.

Why it is required: Duplicating page objects per platform doubles maintenance for no benefit.

Example: A checkout button located on both Android and iOS.

java

1public class CartScreen extends BaseScreen {
2    @AndroidFindBy(accessibility = "checkout_button")
3    @iOSXCUITFindBy(iOSNsPredicate = "name == 'checkout_button'")
4    private WebElement checkout;
5 
6    public CheckoutScreen checkout() {
7        tap(checkout);
8        return new CheckoutScreen(driver);
9    }
10}

Expected result

The same test class runs on Android and iOS drivers.

Common mistakes

XPath-heavy locators (very slow on mobile)
No accessibility ids requested from developers
Testing only on emulators

How do you keep one mobile suite working on both platforms?

Related interview questions

Open bank

How do you handle element location differences between Android and iOS?
Keep a single page object interface with platform-specific locator strategies resolved at runtime (@AndroidFindBy / @iOSXCUITFindBy or a locator map).
Explain the Appium architecture and how you make locators stable.
Appium is a client/server implementing W3C WebDriver; the server delegates to platform drivers (UiAutomator2 on Android, XCUITest on iOS). Stable locators come from accessibility ids set by developers, not XPath.
Explain Appium's architecture.
Your test speaks W3C WebDriver over HTTP to the Appium server, which routes the session to a platform driver (UiAutomator2/Espresso on Android, XCUITest on iOS) that drives the device through the vendor automation framework.
Which capabilities do you set for an Android and an iOS session?
platformName, automationName, deviceName/udid, app or appPackage+appActivity (bundleId on iOS), plus newCommandTimeout, autoGrantPermissions and noReset/fullReset to control state.
Which locator strategies are most reliable on mobile?
accessibility id first (content-desc on Android, accessibilityIdentifier on iOS), then id/resource-id, then UiSelector or iOS class chain/predicate — XPath last.
How do you perform gestures like swipe, scroll and long press?
Build W3C actions with a PointerInput sequence, or call the mobile: scripts (mobile: swipeGesture, mobile: longClickGesture) which are simpler and more reliable on Android.