Showing posts with label How-to. Show all posts
Showing posts with label How-to. Show all posts

Back and other hard keys: three stories

Android 2.0 introduces new behavior and support for handling hard keys such as BACK and MENU, including some special features to support the virtual hard keys that are appearing on recent devices such as Droid.

This article will give you three stories on these changes: from the most simple to the gory details. Pick the one you prefer.

Story 1: Making things easier for developers

If you were to survey the base applications in the Android platform, you would notice a fairly common pattern: add a little bit of magic to intercept the BACK key and do something different. To do this right, the magic needs to look something like this:

@Override
public boolean onKeyDown(int keyCode, KeyEvent event) {
if (keyCode == KeyEvent.KEYCODE_BACK && event.getRepeatCount() == 0) {
// do something on back.
return true;
}

return super.onKeyDown(keyCode, event);
}

How to intercept the BACK key in an Activity is also one of the common questions we see developers ask, so as of 2.0 we have a new little API to make this more simple and easier to discover and get right:

@Override
public void onBackPressed() {
// do something on back.
return;
}

If this is all you care about doing, and you're not worried about supporting versions of the platform before 2.0, then you can stop here. Otherwise, read on.

Story 2: Embracing long press

One of the fairly late addition to the Android platform was the use of long press on hard keys to perform alternative actions. In 1.0 this was long press on HOME for the recent apps switcher and long press on CALL for the voice dialer. In 1.1 we introduced long press on SEARCH for voice search, and 1.5 introduced long press on MENU to force the soft keyboard to be displayed as a backwards compatibility feature for applications that were not yet IME-aware.

(As an aside: long press on MENU was only intended for backwards compatibility, and thus has some perhaps surprising behavior in how strongly the soft keyboard stays up when it is used. This is not intended to be a standard way to access the soft keyboards, and all apps written today should have a more standard and visible way to bring up the IME if they need it.)

Unfortunately the evolution of this feature resulted in a less than optimal implementation: all of the long press detection was implemented in the client-side framework's default key handling code, using timed messages. This resulted in a lot of duplication of code and some behavior problems; since the actual event dispatching code had no concept of long presses and all timing for them was done on the main thread of the application, the application could be slow enough to not update within the long press timeout.

In Android 2.0 this all changes, with a real KeyEvent API and callback functions for long presses. These greatly simplify long press handling for applications, and allow them to interact correctly with the framework. For example: you can override Activity.onKeyLongPress() to supply your own action for a long press on one of the hard keys, overriding the default action provided by the framework.

Perhaps most significant for developers is a corresponding change in the semantics of the BACK key. Previously the default key handling executed the action for this key when it was pressed, unlike the other hard keys. In 2.0 the BACK key is now execute on key up. However, for existing apps, the framework will continue to execute the action on key down for compatibility reasons. To enable the new behavior in your app you must set android:targetSdkVersion in your manifest to 5 or greater.

Here is an example of code an Activity subclass can use to implement special actions for a long press and short press of the CALL key:

@Override
public boolean onKeyLongPress(int keyCode, KeyEvent event) {
if (keyCode == KeyEvent.KEYCODE_CALL) {
// a long press of the call key.
// do our work, returning true to consume it. by
// returning true, the framework knows an action has
// been performed on the long press, so will set the
// canceled flag for the following up event.
return true;
}
return super.onKeyLongPress(keyCode, event);
}
@Override
public boolean onKeyUp(int keyCode, KeyEvent event) {
if (keyCode == KeyEvent.KEYCODE_CALL && event.isTracking()
&& !event.isCanceled()) {
// if the call key is being released, AND we are tracking
// it from an initial key down, AND it is not canceled,
// then handle it.
return true;
}
return super.onKeyUp(keyCode, event);
}

Note that the above code assumes we are implementing different behavior for a key that is normally processed by the framework. If you want to implement long presses for another key, you will also need to override onKeyDown to have the framework track it:

@Override
public boolean onKeyDown(int keyCode, KeyEvent event) {
if (keyCode == KeyEvent.KEYCODE_0) {
// this tells the framework to start tracking for
// a long press and eventual key up. it will only
// do so if this is the first down (not a repeat).
event.startTracking();
return true;
}
return super.onKeyDown(keyCode, event);
}

Story 3: Making a mess with virtual keys

Now we come to the story of our original motivation for all of these changes: support for virtual hard keys, as seen on the Droid and other upcoming devices. Instead of physical buttons, these devices have a touch sensor that extends outside of the visible screen, creating an area for the "hard" keys to live as touch sensitive areas. The low-level input system looks for touches on the screen in this area, and turns these into "virtual" hard key events as appropriate.

To applications these basically look like real hard keys, though the generated events will have a new FLAG_VIRTUAL_HARD_KEY bit set to identify them. Regardless of that flag, in nearly all cases an application can handle these "hard" key events in the same way it has always done for real hard keys.

However, these keys introduce some wrinkles in user interaction. Most important is that the keys exist on the same surface as the rest of the user interface, and they can be easily pressed with the same kind of touches. This can become an issue, for example, when the virtual keys are along the bottom of the screen: a common gesture is to swipe up the screen for scrolling, and it can be very easy to accidentally touch a virtual key at the bottom when doing this.

The solution for this in 2.0 is to introduce a concept of a "canceled" key event. We've already seen this in the previous story, where handling a long press would cancel the following up event. In a similar way, moving from a virtual key press on to the screen will cause the virtual key to be canceled when it goes up.

In fact the previous code already takes care of this — by checking isCanceled() on the key up, canceled virtual keys and long presses will be ignored. There are also individual flags for these two cases, but they should rarely be used by applications and always with the understanding that in the future there may be more reasons for a key event to be canceled.

For existing application, where BACK key compatibility is turned on to execute the action on down, there is still the problem of accidentally detecting a back press when intending to perform a swipe. Though there is no solution for this except to update an application to specify it targets SDK version 5 or later, fortunately the back key is generally positioned on a far side of the virtual key area, so the user is much less likely to accidentally hit it than some of the other keys.

Writing an application that works well on pre-2.0 as well as 2.0 and later versions of the platform is also fairly easy for most common cases. For example, here is code that allows you to handle the back key in an activity correctly on all versions of the platform:

@Override
public boolean onKeyDown(int keyCode, KeyEvent event) {
if (android.os.Build.VERSION.SDK_INT < android.os.Build.VERSION_CODES.ECLAIR
&& keyCode == KeyEvent.KEYCODE_BACK
&& event.getRepeatCount() == 0) {
// Take care of calling this method on earlier versions of
// the platform where it doesn't exist.
onBackPressed();
}

return super.onKeyDown(keyCode, event);
}

@Override
public void onBackPressed() {
// This will be called either automatically for you on 2.0
// or later, or by the code above on earlier versions of the
// platform.
return;
}

For the hard core: correctly dispatching events

One final topic that is worth covering is how to correctly handle events in the raw dispatch functions such as onDispatchEvent() or onPreIme(). These require a little more care, since you can't rely on some of the help the framework provides when it calls the higher-level functions such as onKeyDown(). The code below shows how you can intercept the dispatching of the BACK key such that you correctly execute your action when it is release.

@Override
public boolean dispatchKeyEvent(KeyEvent event) {
if (event.getKeyCode() == KeyEvent.KEYCODE_BACK) {
if (event.getAction() == KeyEvent.ACTION_DOWN
&& event.getRepeatCount() == 0) {

// Tell the framework to start tracking this event.
getKeyDispatcherState().startTracking(event, this);
return true;

} else if (event.getAction() == KeyEvent.ACTION_UP) {
getKeyDispatcherState().handleUpEvent(event);
if (event.isTracking() && !event.isCanceled()) {

// DO BACK ACTION HERE
return true;

}
}
return super.dispatchKeyEvent(event);
} else {
return super.dispatchKeyEvent(event);
}
}

The call to getKeyDispatcherState() returns an object that is used to track the current key state in your window. It is generally available on the View class, and an Activity can use any of its views to retrieve the object if needed.

Integrating Application with Intents

Written in collaboration with Michael Burton, Mob.ly; Ivan Mitrovic, uLocate; and Josh Garnier, OpenTable.

OpenTable, uLocate, and Mob.ly worked together to create a great user experience on Android. We saw an opportunity to enable WHERE and GoodFood users to make reservations on OpenTable easily and seamlessly. This is a situation where everyone wins — OpenTable gets more traffic, WHERE and GoodFood gain functionality to make their applications stickier, and users benefit because they can make reservations with only a few taps of a finger. We were able to achieve this deep integration between our applications by using Android's Intent mechanism. Intents are perhaps one of Android's coolest, most unique, and under-appreciated features. Here's how we exploited them to compose a new user experience from parts each of us have.

Designing

One of the first steps is to design your Intent interface, or API. The main public Intent that OpenTable exposes is the RESERVE Intent, which lets you make a reservation at a specific restaurant and optionally specify the date, time, and party size.

Hereʼs an example of how to make a reservation using the RESERVE Intent:

startActivity(new Intent("com.opentable.action.RESERVE",
Uri.parse("reserve://opentable.com/2947?partySize=3")));

Our objective was to make it simple and clear to the developer using the Intent. So how did we decide what it would look like?

First, we needed an Action. We considered using Intent.ACTION_VIEW, but decided this didn't map well to making a reservation, so we made up a new action. Following the conventions of the Android platform (roughly <package-name>.action.<action-name>), we chose "com.opentable.action.RESERVE". Actions really are just strings, so it's important to namespace them. Not all applications will need to define their own actions. In fact, common actions such as Intent.ACTION_VIEW (aka "android.intent.action.VIEW") are often a better choice if youʼre not doing something unusual.

Next we needed to determine how data would be sent in our Intent. We decided to have the data encoded in a URI, although you might choose to receive your data as a collection of items in the Intent's data Bundle. We used a scheme of "reserve:" to be consistent with our action. We then put our domain authority and the restaurant ID into the URI path since it was required, and we shunted off all of the other, optional inputs to URI query parameters.

Exposing

Once we knew what we wanted the Intent to look like, we needed to register the Intent with the system so Android would know to start up the OpenTable application. This is done by inserting an Intent filter into the appropriate Activity declaration in AndroidManifest.xml:

<activity android:name=".activity.Splash" ... >
...
<intent-filter>
<action android:name="com.opentable.action.RESERVE"/>
<category android:name="android.intent.category.DEFAULT" />
<data android:scheme="reserve" android:host="opentable.com"/>
</intent-filter>
...
</activity>

In our case, we wanted users to see a brief OpenTable splash screen as we loaded up details about their restaurant selection, so we put the Intent Filter in the splash Activity definition. We set our category to be DEFAULT. This will ensure our application is launched without asking the user what application to use, as long as no other Activities also list themselves as default for this action.

Notice that things like the URI query parameter ("partySize" in our example) are not specified by the Intent filter. This is why documentation is key when defining your Intents, which weʼll talk about a bit later.

Processing

Now the only thing left to do was write the code to handle the intent.

    protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
final Uri uri;
final int restaurantId;
try {
uri = getIntent().getData();
restaurantId = Integer.parseInt( uri.getPathSegments().get(0));
} catch(Exception e) {
// Restaurant ID is required
Log.e(e);
startActivity( FindTable.start(FindTablePublic.this));
finish();
return;
}
final String partySize = uri.getQueryParameter("partySize");
...
}

Although this is not quite all the code, you get the idea. The hardest part here was the error handling. OpenTable wanted to be able to gracefully handle erroneous Intents that might be sent by partner applications, so if we have any problem parsing the restaurant ID, we pass the user off to another Activity where they can find the restaurant manually. It's important to verify the input just as you would in a desktop or web application to protect against injection attacks that might harm your app or your users.

Calling and Handling Uncertainty with Grace

Actually invoking the target application from within the requester is quite straight-forward, but there are a few cases we need to handle. What if OpenTable isn't installed? What if WHERE or GoodFood doesn't know the restaurant ID?



Restaurant ID knownRestaurant ID unknown
User has OpenTableCall OpenTable IntentDon't show reserve button
User doesn't have OpenTableCall Market IntentDon't show reserve button

You'll probably wish to work with your partner to decide exactly what to do if the user doesn't have the target application installed. In this case, we decided we would take the user to Android Market to download OpenTable if s/he wished to do so.

    public void showReserveButton() {

// setup the Intent to call OpenTable
Uri reserveUri = Uri.parse(String.format( "reserve://opentable.com/%s?refId=5449",
opentableId));
Intent opentableIntent = new Intent("com.opentable.action.RESERVE", reserveUri);

// setup the Intent to deep link into Android Market
Uri marketUri = Uri.parse("market://search?q=pname:com.opentable");
Intent marketIntent = new Intent(Intent.ACTION_VIEW).setData(marketUri);

opentableButton.setVisibility(opentableId > 0 ? View.VISIBLE : View.GONE);
opentableButton.setOnClickListener(new Button.OnClickListener() {
public void onClick(View v) {
PackageManager pm = getPackageManager();
startActivity(pm.queryIntentActivities(opentableIntent, 0).size() == 0 ?
opentableIntent : marketIntent);
}
});
}

In the case where the ID for the restaurant is unavailable, whether because they don't take reservations or they aren't part of the OpenTable network, we simply hide the reserve button.



Publishing the Intent Specification

Now that all the technical work is done, how can you get other developers to use your Intent-based API besides 1:1 outreach? The answer is simple: publish documentation on your website. This makes it more likely that other applications will link to your functionality and also makes your application available to a wider community than you might otherwise reach.

If there's an application that you'd like to tap into that doesn't have any published information, try contacting the developer. It's often in their best interest to encourage third parties to use their APIs, and if they already have an API sitting around, it might be simple to get you the documentation for it.

Summary

It's really just this simple. Now when any of us is in a new city or just around the neighborhood its easy to check which place is the new hot spot and immediately grab an available table. Its great to not need to find a restaurant in one application, launch OpenTable to see if there's a table, find out there isn't, launch the first application again, and on and on. We hope you'll find this write-up useful as you develop your own public intents and that you'll consider sharing them with the greater Android community.

UI framework changes in Android 1.6

Android 1.6 introduces numerous enhancements and bug fixes in the UI framework. Today, I'd like to highlight three two improvements in particular.

Optimized drawing

The UI toolkit introduced in Android 1.6 is aware of which views are opaque and can use this information to avoid drawing views that the user will not be able to see. Before Android 1.6, the UI toolkit would sometimes perform unnecessary operations by drawing a window background when it was obscured by a full-screen opaque view. A workaround was available to avoid this, but the technique was limited and required work on your part. With Android 1.6, the UI toolkit determines whether a view is opaque by simply querying the opacity of the background drawable. If you know that your view is going to be opaque but that information does not depend on the background drawable, you can simply override the method called isOpaque():

@Override
public boolean isOpaque() {
return true;
}

The value returned by isOpaque() does not have to be constant and can change at any time. For instance, the implementation of ListView in Android 1.6 indicates that a list is opaque only when the user is scrolling it.

Updated: Our apologies—we spoke to soon about isOpaque(). It will be available in a future update to the Android platform.

More flexible, more robust RelativeLayout

RelativeLayout is the most versatile layout offered by the Android UI toolkit and can be successfully used to reduce the number of views created by your applications. This layout used to suffer from various bugs and limitations, sometimes making it difficult to use without having some knowledge of its implementation. To make your life easier, Android 1.6 comes with a revamped RelativeLayout. This new implementation not only fixes all known bugs in RelativeLayout (let us know when you find new ones) but also addresses its major limitation: the fact that views had to be declared in a particular order. Consider the following XML layout:

<?xml version="1.0" encoding="utf-8"?>

<RelativeLayout xmlns:android="http://schemas.android.com/apk/res/android"
android:layout_width="fill_parent"
android:layout_height="64dip"
android:padding="6dip">

<TextView
android:id="@+id/band"
android:layout_width="fill_parent"
android:layout_height="26dip"

android:layout_below="@+id/track"
android:layout_alignLeft="@id/track"
android:layout_alignParentBottom="true"

android:gravity="top"
android:text="The Airborne Toxic Event" />

<TextView
android:id="@id/track"
android:layout_marginLeft="6dip"
android:layout_width="fill_parent"
android:layout_height="26dip"

android:layout_toRightOf="@+id/artwork"

android:textAppearance="?android:attr/textAppearanceMedium"
android:gravity="bottom"
android:text="Sometime Around Midnight" />

<ImageView
android:id="@id/artwork"
android:layout_width="56dip"
android:layout_height="56dip"
android:layout_gravity="center_vertical"

android:src="@drawable/artwork" />

</RelativeLayout>

This code builds a very simple layout—an image on the left with two lines of text stacked vertically. This XML layout is perfectly fine and contains no errors. Unfortunately, Android 1.5's RelativeLayout is incapable of rendering it correctly, as shown in the screenshot below.

The problem is that this layout uses forward references. For instance, the "band" TextView is positioned below the "track" TextView but "track" is declared after "band" and, in Android 1.5, RelativeLayout does not know how to handle this case. Now look at the exact same layout running on Android 1.6:

As you can see Android 1.6 is now better able to handle forward reference. The result on screen is exactly what you would expect when writing the layout.

Easier click listeners

Setting up a click listener on a button is very common task, but it requires quite a bit of boilerplate code:

findViewById(R.id.myButton).setOnClickListener(new View.OnClickListener() {
public void onClick(View v) {
// Do stuff
}
});

One way to reduce the amount of boilerplate is to share a single click listener between several buttons. While this technique reduces the number of classes, it still requires a fair amount of code and it still requires giving each button an id in your XML layout file:

View.OnClickListener handler = View.OnClickListener() {
public void onClick(View v) {
switch (v.getId()) {
case R.id.myButton: // doStuff
break;
case R.id.myOtherButton: // doStuff
break;
}
}
}

findViewById(R.id.myButton).setOnClickListener(handler);
findViewById(R.id.myOtherButton).setOnClickListener(handler);

With Android 1.6, none of this is necessary. All you have to do is declare a public method in your Activity to handle the click (the method must have one View argument):

class MyActivity extends Activity {
public void myClickHandler(View target) {
// Do stuff
}
}

And then reference this method from your XML layout:

<Button android:onClick="myClickHandler" />

This new feature reduces both the amount of Java and XML you have to write, leaving you more time to concentrate on your application.

The Android team is committed to helping you write applications in the easiest and most efficient way possible. We hope you find these improvements useful and we're excited to see your applications on Android Market.

Gestures on Android 1.6

Touch screens are a great way to interact with applications on mobile devices. With a touch screen, users can easily tap, drag, fling, or slide to quickly perform actions in their favorite applications. But it's not always that easy for developers. With Android, it's easy to recognize simple actions, like a swipe, but it's much more difficult to handle complicated gestures, which also require developers to write a lot of code. That's why we have decided to introduce a new gestures API in Android 1.6. This API, located in the new package android.gesture, lets you store, load, draw and recognize gestures. In this post I will show you how you can use the android.gesture API in your applications. Before going any further, you should download the source code of the examples.

Creating a gestures library

The Android 1.6 SDK comes with a new application pre-installed on the emulator, called Gestures Builder. You can use this application to create a set of pre-defined gestures for your own application. It also serves as an example of how to let the user define his own gestures in your applications. You can find the source code of Gestures Builders in the samples directory of Android 1.6. In our example we will use Gestures Builder to generate a set of gestures for us (make sure to create an AVD with an SD card image to use Gestures Builder.) The screenshot below shows what the application looks like after adding a few gestures:

As you can see, a gesture is always associated with a name. That name is very important because it identifies each gesture within your application. The names do not have to be unique. Actually it can be very useful to have several gestures with the same name to increase the precision of the recognition. Every time you add or edit a gesture in the Gestures Builder, a file is generated on the emulator's SD card, /sdcard/gestures. This file contains the description of all the gestures, and you will need to package it inside your application inside the resources directory, in /res/raw.

Loading the gestures library

Now that you have a set of pre-defined gestures, you must load it inside your application. This can be achieved in several ways but the easiest is to use the GestureLibraries class:

mLibrary = GestureLibraries.fromRawResource(this, R.raw.spells);
if (!mLibrary.load()) {
finish();
}

In this example, the gesture library is loaded from the file /res/raw/spells. You can easily load libraries from other sources, like the SD card, which is very important if you want your application to be able to save the library; a library loaded from a raw resource is read-only and cannot be modified. The following diagram shows the structure of a library:

Recognizing gestures

To start recognizing gestures in your application, all you have to do is add a GestureOverlayView to your XML layout:

<android.gesture.GestureOverlayView
android:id="@+id/gestures"
android:layout_width="fill_parent"
android:layout_height="0dip"
android:layout_weight="1.0" />

Notice that the GestureOverlayView is not part of the usual android.widget package. Therefore, you must use its fully qualified name. A gesture overlay acts as a simple drawing board on which the user can draw his gestures. You can tweak several visual properties, like the color and the width of the stroke used to draw gestures, and register various listeners to follow what the user is doing. The most commonly used listener is GestureOverlayView.OnGesturePerformedListener which fires whenever a user is done drawing a gesture:

GestureOverlayView gestures = (GestureOverlayView) findViewById(R.id.gestures);
gestures.addOnGesturePerformedListener(this);

When the listener fires, you can ask the GestureLibrary to try to recognize the gesture. In return, you will get a list of Prediction instances, each with a name - the same name you entered in the Gestures Builder - and a score. The list is sorted by descending scores; the higher the score, the more likely the associated gesture is the one the user intended to draw. The following code snippet demonstrates how to retrieve the name of the first prediction:

public void onGesturePerformed(GestureOverlayView overlay, Gesture gesture) {
ArrayList predictions = mLibrary.recognize(gesture);

// We want at least one prediction
if (predictions.size() > 0) {
Prediction prediction = predictions.get(0);
// We want at least some confidence in the result
if (prediction.score > 1.0) {
// Show the spell
Toast.makeText(this, prediction.name, Toast.LENGTH_SHORT).show();
}
}
}

In this example, the first prediction is taken into account only if it's score is greater than 1.0. The threshold you use is entirely up to you but know that scores lower than 1.0 are typically poor matches. And this is all the code you need to create a simple application that can recognize pre-defined gestures (see the source code of the project GesturesDemo):

Gestures overlay

In the example above, the GestureOverlayView was used as a normal view, embedded inside a LinearLayout. However, as its name suggests, it can also be used as an overlay on top of other views. This can be useful to recognize gestures in a game or just anywhere in the UI of an application. In the second example, called GesturesListDemo, we'll create an overlay on top of a list of contacts. We start again in Gestures Builder to create a new set of pre-defined gestures:

And here is what the XML layout looks like:

<android.gesture.GestureOverlayView
xmlns:android="http://schemas.android.com/apk/res/android"
android:id="@+id/gestures"
android:layout_width="fill_parent"
android:layout_height="fill_parent"

android:gestureStrokeType="multiple"
android:eventsInterceptionEnabled="true"
android:orientation="vertical">

<ListView
android:id="@android:id/list"
android:layout_width="fill_parent"
android:layout_height="fill_parent" />

</android.gesture.GestureOverlayView>

In this application, the gestures view is an overlay on top of a regular ListView. The overlay also specifies a few properties that we did not need before:

  • gestureStrokeType: indicates whether we want to recognize gestures made of a single stroke or multiple strokes. Since one of our gestures is the "+" symbol, we need multiple strokes
  • eventsInterceptionEnabled: when set to true, this property tells the overlay to steal the events from its children as soon as it knows the user is really drawing a gesture. This is useful when there's a scrollable view under the overlay, to avoid scrolling the underlying child as the user draws his gesture
  • orientation: indicates the scroll orientation of the views underneath. In this case the list scrolls vertically, which means that any horizontal gestures (like action_delete) can immediately be recognized as a gesture. Gestures that start with a vertical stroke must contain at least one horizontal component to be recognized. In other words, a simple vertical line cannot be recognized as a gesture since it would conflict with the list's scrolling.

The code used to load and set up the gestures library and overlay is exactly the same as before. The only difference is that we now check the name of the predictions to know what the user intended to do:

public void onGesturePerformed(GestureOverlayView overlay, Gesture gesture) {
ArrayList<Prediction> predictions = mLibrary.recognize(gesture);
if (predictions.size() > 0 && predictions.get(0).score > 1.0) {
String action = predictions.get(0).name;
if ("action_add".equals(action)) {
Toast.makeText(this, "Adding a contact", Toast.LENGTH_SHORT).show();
} else if ("action_delete".equals(action)) {
Toast.makeText(this, "Removing a contact", Toast.LENGTH_SHORT).show();
} else if ("action_refresh".equals(action)) {
Toast.makeText(this, "Reloading contacts", Toast.LENGTH_SHORT).show();
}
}
}

The user is now able to draw his gestures on top of the list without interfering with the scrolling:

The overlay even gives visual clues as to whether the gesture is considered valid for recognition. In the case of a vertical overlay, for instance, a single vertical stroke cannot be recognized as a gesture and is therefore drawn with a translucent color:

It's your turn

Adding support for gestures in your application is easy and can be a valuable addition. The gestures API does not even have to be used to recognize complex shapes; it will work equally well to recognize simple swipes. We are very excited by the possibilities the gestures API offers, and we're eager to see what cool applications the community will create with it.

Zipalign: an easy optimization

The Android 1.6 SDK includes a tool called zipalign that optimizes the way an application is packaged. Doing this enables Android to interact with your application more efficiently and thus has the potential to make your application and the overall system run faster. We strongly encourage you to use zipalign on both new and already published applications and to make the optimized version available—even if your application targets a previous version of Android. We'll get into more detail on what zipalign does, how to use it, and why you'll want to do so in the rest of this post.

In Android, data files stored in each application's apk are accessed by multiple processes: the installer reads the manifest to handle the permissions associated with that application; the Home application reads resources to get the application's name and icon; the system server reads resources for a variety of reasons (e.g. to display that application's notifications); and last but not least, the resource files are obviously used by the application itself.

The resource-handling code in Android can efficiently access resources when they're aligned on 4-byte boundaries by memory-mapping them. But for resources that are not aligned (i.e. when zipalign hasn't been run on an apk), it has to fall back to explicitly reading them—which is slower and consumes additional memory.

For an application developer like you, this fallback mechanism is very convenient. It provides a lot of flexibility by allowing for several different development methods, including those that don't include aligning resources as part of their normal flow.

Unfortunately, the situation is reversed for users—reading resources from unaligned apks is slow and takes a lot of memory. In the best case, the only visible result is that both the Home application and the unaligned application launch slower than they otherwise should. In the worst case, installing several applications with unaligned resources increases memory pressure, thus causing the system to thrash around by having to constantly start and kill processes. The user ends up with a slow device with a poor battery life.

Luckily, it's very easy to align the resources:

  • Using ADT:
    • ADT (starting with 0.9.3) will automatically align release application packages if the export wizard is used to create them. To use the wizard, right click the project and choose "Android Tools" > "Export Signed Application Package..." It can also be accessed from the first page of the AndroidManifest.xml editor.
  • Using Ant:
    • The Ant build script that targets Android 1.6 (API level 4) can align application packages. Targets for older versions of the Android platform are not aligned by the Ant build script and need to be manually aligned.
    • Debug packages built with Ant for Android 1.6 applications are aligned and signed by default.
    • Release packages are aligned automatically only if Ant has enough information to sign the packages, since aligning has to happen after signing. In order to be able to sign packages, and therefore to align them, Ant needs to know the location of the keystore and the name of the key in build.properties. The name of the properties are key.store and key.alias respectively. If those properties are present, the signing tool will prompt to enter the store/key passwords during the build, and the script will sign and then align the apk file. If the properties are missing, the release package will not be signed, and therefore will not get aligned either.
  • Manually:
    • In order to manually align a package, zipalign is in the tools folder of the Android 1.6 SDK. It can be used on application packages targeting any version of Android. It should be run after signing the apk file, using the following command:
      zipalign -v 4 source.apk destination.apk
  • Verifying alignment:
    • The following command verifies that a package is aligned:
      zipalign -c -v 4 application.apk

We encourage you manually run zipalign on your currently published applications and to make the newly aligned versions available to users. And don't forget to align any new applications going forward!

Introducing Quick Search Box for Android

One of the new features we're really proud of in the Android 1.6 release is Quick Search Box for Android. This is our new system-wide search framework, which makes it possible for users to quickly and easily find what they're looking for, both on their devices and on the web. It suggests content on your device as you type, like apps, contacts, browser history, and music. It also offers results from the web search suggestions, local business listings, and other info from Google, such as stock quotes, weather, and flight status. All of this is available right from the home screen, by tapping on Quick Search Box (QSB).

What we're most excited about with this new feature is the ability for you, the developers, to leverage the QSB framework to provide quicker and easier access to the content inside your apps. Your apps can provide search suggestions that will surface to users in QSB alongside other search results and suggestions. This makes it possible for users to access your application's content from outside your application—for example, from the home screen.

The code fragments below are related to a new demo app for Android 1.6 called Searchable Dictionary.


The story before now: searching within your app

In previous releases, we already provided a mechanism for you to expose search and search suggestions in your app as described in the docs for SearchManager. This mechanism has not changed and requires the following two things in your AndroidManifest.xml:

1) In your <activity>, an intent filter, and a reference to a searchable.xml file (described below):

<intent-filter>
<action android:name="android.intent.action.SEARCH" />
<category android:name="android.intent.category.DEFAULT" />
</intent-filter>

<meta-data android:name="android.app.searchable"
android:resource="@xml/searchable" />

2) A content provider that can provide search suggestions according to the URIs and column formats specified by the Search Suggestions section of the SearchManager docs:

<!-- Provides search suggestions for words and their definitions. -->
<provider android:name="DictionaryProvider"
android:authorities="dictionary"
android:syncable="false" />

In the searchable.xml file, you specify a few things about how you want the search system to present search for your app, including the authority of the content provider that provides suggestions for the user as they type. Here's an example of the searchable.xml of an Android app that provides search suggestions within its own activities:

<searchable xmlns:android="http://schemas.android.com/apk/res/android"
android:label="@string/search_label"
android:searchSuggestAuthority="dictionary"
android:searchSuggestIntentAction="android.intent.action.VIEW">
</searchable>

Note that the android:searchSuggestAuthority attribute refers to the authority of the content provider we declared in AndroidManifest.xml.

For more details on this, see the Searchability Metadata section of the SearchManager docs.

Including your app in Quick Search Box

In Android 1.6, we added a new attribute to the metadata for searchables: android:includeInGlobalSearch. By specifying this as "true" in your searchable.xml, you allow QSB to pick up your search suggestion content provider and include its suggestions along with the rest (if the user enables your suggestions from the system search settings).

You should also specify a string value for android:searchSettingsDescription, which describes to users what sorts of suggestions your app provides in the system settings for search.

<searchable xmlns:android="http://schemas.android.com/apk/res/android"
android:label="@string/search_label"
android:searchSettingsDescription="@string/settings_description"
android:includeInGlobalSearch="true"
android:searchSuggestAuthority="dictionary"
android:searchSuggestIntentAction="android.intent.action.VIEW">
</searchable>

These new attributes are supported only in Android 1.6 and later.

What to expect

The first and most important thing to note is that when a user installs an app with a suggestion provider that participates in QSB, this new app will not be enabled for QSB by default. The user can choose to enable particular suggestion sources from the system settings for search (by going to "Search" > "Searchable items" in settings).

You should consider how to handle this in your app. Perhaps show a notice that instructs the user to visit system settings and enable your app's suggestions.

Once the user enables your searchable item, the app's suggestions will have a chance to show up in QSB, most likely under the "more results" section to begin with. As your app's suggestions are chosen more frequently, they can move up in the list.

Shortcuts

One of our objectives with QSB is to make it faster for users to access the things they access most often. One way we've done this is by 'shortcutting' some of the previously chosen search suggestions, so they will be shown immediately as the user starts typing, instead of waiting to query the content providers. Suggestions from your app may be chosen as shortcuts when the user clicks on them.

For dynamic suggestions that may wish to change their content (or become invalid) in the future, you can provide a 'shortcut id'. This tells QSB to query your suggestion provider for up-to-date content for a suggestion after it has been displayed. For more details on how to manage shortcuts, see the Shortcuts section within the SearchManager docs.


QSB provides a really cool way to make your app's content quicker to access by users. To help you get your app started with it, we've created a demo app which simply provides access to a small dictionary of words in QSB—it's called Searchable Dictionary, and we encourage you to check it out.

Painless threading

Whenever you first start an Android application, a thread called "main" is automatically created. The main thread, also called the UI thread, is very important because it is in charge of dispatching the events to the appropriate widgets and this includes the drawing events. It is also the thread you interact with Android widgets on. For instance, if you touch the a button on screen, the UI thread dispatches the touch event to the widget which in turn sets its pressed state and posts an invalidate request to the event queue. The UI thread dequeues the request and notifies the widget to redraw itself.

This single thread model can yield poor performance in Android applications that do not consider the implications. Since everything happens on a single thread performing long operations, like network access or database queries, on this thread will block the whole user interface. No event can be dispatched, including drawing events, while the long operation is underway. From the user's perspective, the application appears hung. Even worse, if the UI thread is blocked for more than a few seconds (about 5 seconds currently) the user is presented with the infamous "application not responding" (ANR) dialog.

If you want to see how bad this can look, write a simple application with a button that invokes Thread.sleep(2000) in its OnClickListener. The button will remain in its pressed state for about 2 seconds before going back to its normal state. When this happens, it is very easy for the user to perceive the application as slow.

Now that you know you must avoid lengthy operations on the UI thread, you will probably use extra threads (background or worker threads) to perform these operations, and rightly so. Let's take the example of a click listener downloading an image over the network and displaying it in an ImageView:

public void onClick(View v) {
new Thread(new Runnable() {
public void run() {
Bitmap b = loadImageFromNetwork();
mImageView.setImageBitmap(b);
}
}).start();
}

At first, this code seems to be a good solution to your problem, as it does not block the UI thread. Unfortunately, it violates the single thread model: the Android UI toolkit is not thread-safe and must always be manipulated on the UI thread. In this piece of code, the ImageView is manipulated on a worker thread, which can cause really weird problems. Tracking down and fixing such bugs can be difficult and time-consuming.

Android offers several ways to access the UI thread from other threads. You may already be familiar with some of them but here is a comprehensive list:

Any of these classes and methods could be used to correct our previous code example:

public void onClick(View v) {
new Thread(new Runnable() {
public void run() {
final Bitmap b = loadImageFromNetwork();
mImageView.post(new Runnable() {
public void run() {
mImageView.setImageBitmap(b);
}
});
}
}).start();
}

Unfortunately, these classes and methods also tend to make your code more complicated and more difficult to read. It becomes even worse when your implement complex operations that require frequent UI updates. To remedy this problem, Android 1.5 offers a new utility class, called AsyncTask, that simplifies the creation of long-running tasks that need to communicate with the user interface.

AsyncTask is also available for Android 1.0 and 1.1 under the name UserTask. It offers the exact same API and all you have to do is copy its source code in your application.

The goal of AsyncTask is to take care of thread management for you. Our previous example can easily be rewritten with AsyncTask:

public void onClick(View v) {
new DownloadImageTask().execute("http://example.com/image.png");
}

private class DownloadImageTask extends AsyncTask {
protected Bitmap doInBackground(String... urls) {
return loadImageFromNetwork(urls[0]);
}

protected void onPostExecute(Bitmap result) {
mImageView.setImageBitmap(result);
}
}

As you can see, AsyncTask must be used by subclassing it. It is also very important to remember that an AsyncTask instance has to be created on the UI thread and can be executed only once. You can read the AsyncTask documentation for a full understanding on how to use this class, but here is a quick overview of how it works:

In addition to the official documentation, you can read several complex examples in the source code of Shelves (ShelvesActivity.java and AddBookActivity.java) and Photostream (LoginActivity.java, PhotostreamActivity.java and ViewPhotoActivity.java). I highly recommend reading the source code of Shelves to see how to persist tasks across configuration changes and how to cancel them properly when the activity is destroyed.

Regardless of whether or not you use AsyncTask, always remember these two rules about the single thread model: do not block the UI thread and make sure the Android UI toolkit is only accessed on the UI thread. AsyncTask just makes it easier to do both of these things.

If you want to learn more cool techniques, come join us at Google I/O. Members of the Android team will be there to give a series of in-depth technical sessions and answer all your questions.

Drawable mutations

Android's drawables are extremely useful to easily build applications. A Drawable is a pluggable drawing container that is usually associated with a View. For instance, a BitmapDrawable is used to display images, a ShapeDrawable to draw shapes and gradients, etc. You can even combine them to create complex renderings.

Drawables allow you to easily customize the rendering of the widgets without subclassing them. As a matter of fact, they are so convenient that most of the default Android apps and widgets are built using drawables; there are about 700 drawables used in the core Android framework. Because drawables are used so extensively throughout the system, Android optimizes them when they are loaded from resources. For instance, every time you create a Button, a new drawable is loaded from the framework resources (android.R.drawable.btn_default). This means all buttons across all the apps use a different drawable instance as their background. However, all these drawables share a common state, called the "constant state." The content of this state varies according to the type of drawable you are using, but it usually contains all the properties that can be defined by a resource. In the case of a button, the constant state contains a bitmap image. This way, all buttons across all applications share the same bitmap, which saves a lot of memory.

The following diagram shows what entities are created when you assign the same image resource as the background of two different views. As you can see, two drawables are created but they both share the same constant state, hence the same bitmap:

This state sharing feature is great to avoid wasting memory but it can cause problems when you try to modify the properties of a drawable. Imagine an application with a list of books. Each book has a star next to its name, totally opaque when the user marks the book as a favorite, and translucent when the book is not a favorite. To achieve this effect, you would probably write the following code in your list adapter's getView() method:

Book book = ...;
TextView listItem = ...;

listItem.setText(book.getTitle());

Drawable star = context.getResources().getDrawable(R.drawable.star);
if (book.isFavorite()) {
star.setAlpha(255); // opaque
} else {
star.setAlpha(70); // translucent
}

Unfortunately, this piece of code yields a rather strange result, all the drawables have the same opacity:

This result is explained by the constant state. Even though we are getting a new drawable instance for each list item, the constant state remains the same and, in the case of BitmapDrawable, the opacity is part of the constant state. Thus, changing the opacity of one drawable instance changes the opacity of all the other instances. Even worse, working around this issue was not easy with Android 1.0 and 1.1.

Android 1.5 offers a very way to solve this issue with a the new mutate() method. When you invoke this method on a drawable, the constant state of the drawable is duplicated to allow you to change any property without affecting other drawables. Note that bitmaps are still shared, even after mutating a drawable. The diagram below shows what happens when you invoke mutate() on a drawable:

Let's update our previous piece of code to make use of mutate():

Drawable star = context.getResources().getDrawable(R.drawable.star);
if (book.isFavorite()) {
star.mutate().setAlpha(255); // opaque
} else {
star. mutate().setAlpha(70); // translucent
}

For convenience, mutate() returns the drawable itself, which allows to chain method calls. It does not however create a new drawable instance. With this new piece of code, our application now behaves correctly:

If you want to learn more cool techniques, come join us at Google I/O. Members of the Android team will be there to give a series of in-depth technical sessions and answer all your questions.

Backward compatibility for Android applications

Android 1.5 introduced a number of new features that application developers can take advantage of, like virtual input devices and speech recognition. As a developer, you need to be aware of backward compatibility issues on older devices—do you want to allow your application to run on all devices, or just those running newer software? In some cases it will be useful to employ the newer APIs on devices that support them, while continuing to support older devices.

If the use of a new API is integral to the program—perhaps you need to record video—you should add a manifest entry to ensure your app won't be installed on older devices. For example, if you require APIs added in 1.5, you would specify 3 as the minimum SDK version:

  <manifest>
...
<uses-sdk android:minSdkVersion="3" />
...
</manifest>

If you want to add a useful but non-essential feature, such as popping up an on-screen keyboard even when a hardware keyboard is available, you can write your program in a way that allows it to use the newer features without failing on older devices.

Using reflection

Suppose there's a simple new call you want to use, like android.os.Debug.dumpHprofData(String filename). The android.os.Debug class has existed since the first SDK, but the method is new in 1.5. If you try to call it directly, your app will fail to run on older devices.

The simplest way to call the method is through reflection. This requires doing a one-time lookup and caching the result in a Method object. Using the method is a matter of calling Method.invoke and un-boxing the result. Consider the following:

public class Reflect {
private static Method mDebug_dumpHprofData;

static {
initCompatibility();
};

private static void initCompatibility() {
try {
mDebug_dumpHprofData = Debug.class.getMethod(
"dumpHprofData", new Class[] { String.class } );
/* success, this is a newer device */
} catch (NoSuchMethodException nsme) {
/* failure, must be older device */
}
}

private static void dumpHprofData(String fileName) throws IOException {
try {
mDebug_dumpHprofData.invoke(null, fileName);
} catch (InvocationTargetException ite) {
/* unpack original exception when possible */
Throwable cause = ite.getCause();
if (cause instanceof IOException) {
throw (IOException) cause;
} else if (cause instanceof RuntimeException) {
throw (RuntimeException) cause;
} else if (cause instanceof Error) {
throw (Error) cause;
} else {
/* unexpected checked exception; wrap and re-throw */
throw new RuntimeException(ite);
}
} catch (IllegalAccessException ie) {
System.err.println("unexpected " + ie);
}
}

public void fiddle() {
if (mDebug_dumpHprofData != null) {
/* feature is supported */
try {
dumpHprofData("/sdcard/dump.hprof");
} catch (IOException ie) {
System.err.println("dump failed!");
}
} else {
/* feature not supported, do something else */
System.out.println("dump not supported");
}
}
}

This uses a static initializer to call initCompatibility, which does the method lookup. If that succeeds, it uses a private method with the same semantics as the original (arguments, return value, checked exceptions) to do the call. The return value (if it had one) and exception are unpacked and returned in a way that mimics the original. The fiddle method demonstrates how the application logic would choose to call the new API or do something different based on the presence of the new method.

For each additional method you want to call, you would add an additional private Method field, field initializer, and call wrapper to the class.

This approach becomes a bit more complex when the method is declared in a previously undefined class. It's also much slower to call Method.invoke() than it is to call the method directly. These issues can be mitigated by using a wrapper class.

Using a wrapper class

The idea is to create a class that wraps all of the new APIs exposed by a new or existing class. Each method in the wrapper class just calls through to the corresponding real method and returns the same result.

If the target class and method exist, you get the same behavior you would get by calling the class directly, with a small amount of overhead from the additional method call. If the target class or method doesn't exist, the initialization of the wrapper class fails, and your application knows that it should avoid using the newer calls.

Suppose this new class were added:

public class NewClass {
private static int mDiv = 1;

private int mMult;

public static void setGlobalDiv(int div) {
mDiv = div;
}

public NewClass(int mult) {
mMult = mult;
}

public int doStuff(int val) {
return (val * mMult) / mDiv;
}
}

We would create a wrapper class for it:

class WrapNewClass {
private NewClass mInstance;

/* class initialization fails when this throws an exception */
static {
try {
Class.forName("NewClass");
} catch (Exception ex) {
throw new RuntimeException(ex);
}
}

/* calling here forces class initialization */
public static void checkAvailable() {}

public static void setGlobalDiv(int div) {
NewClass.setGlobalDiv(div);
}

public WrapNewClass(int mult) {
mInstance = new NewClass(mult);
}

public int doStuff(int val) {
return mInstance.doStuff(val);
}
}

This has one method for each constructor and method in the original, plus a static initializer that tests for the presence of the new class. If the new class isn't available, initialization of WrapNewClass fails, ensuring that the wrapper class can't be used inadvertently. The checkAvailable method is used as a simple way to force class initialization. We use it like this:

public class MyApp {
private static boolean mNewClassAvailable;

/* establish whether the "new" class is available to us */
static {
try {
WrapNewClass.checkAvailable();
mNewClassAvailable = true;
} catch (Throwable t) {
mNewClassAvailable = false;
}
}

public void diddle() {
if (mNewClassAvailable) {
WrapNewClass.setGlobalDiv(4);
WrapNewClass wnc = new WrapNewClass(40);
System.out.println("newer API is available - " + wnc.doStuff(10));
} else {
System.out.println("newer API not available");
}
}
}

If the call to checkAvailable succeeds, we know the new class is part of the system. If it fails, we know the class isn't there, and adjust our expectations accordingly. It should be noted that the call to checkAvailable will fail before it even starts if the bytecode verifier decides that it doesn't want to accept a class that has references to a nonexistent class. The way this code is structured, the end result is the same whether the exception comes from the verifier or from the call to Class.forName.

When wrapping an existing class that now has new methods, you only need to put the new methods in the wrapper class. Invoke the old methods directly. The static initializer in WrapNewClass would be augmented to do a one-time check with reflection.

Testing is key

You must test your application on every version of the Android framework that is expected to support it. By definition, the behavior of your application will be different on each. Remember the mantra: if you haven't tried it, it doesn't work.

You can test for backward compatibility by running your application in an emulator from an older SDK, but as of the 1.5 release there's a better way. The SDK allows you to specify "Android Virtual Devices" with different API levels. Once you create the AVDs, you can test your application with old and new versions of the system, perhaps running them side-by-side to see the differences. More information about emulator AVDs can be found in the SDK documentation and from emulator -help-virtual-device.


Learn about Android 1.5 and more at Google I/O. Members of the Android team will be there to give a series of in-depth technical sessions and to field your toughest questions.

Future-Proofing Your Apps

Hi, developers! I hope you've heard about the early-look version of the Android 1.5 SDK that we recently released. There are some great new features in there, but don't get too excited yet -- some of you will need to fix some problems in your apps before you can start taking advantage of Android 1.5.

We've done some fairly extensive testing of the popular apps on the Android Market, and it turns out that a few of those apps use some bad techniques that cause them to crash or behave strangely on Android 1.5. The list below is based on our observations of five ways that we've seen bad apps fail on 1.5. You can think of these as "anti-patterns" (that is, techniques to avoid) for Android development. If you've written an app with the Android 1.0 or 1.1 SDKs, you'll need to pay close attention.

Technique to Avoid, #1: Using Internal APIs

Even though we've always strongly advised against doing so, some developers have chosen to use unsupported or internal APIs. For instance, many developers are using the internal brightness control and bluetooth toggle APIs that were present in 1.0 and 1.1. A bug -- which is now fixed in Android 1.5 -- allowed apps to use those APIs without requesting permission. As a result, apps that use those APIs will break on 1.5. There are other changes to unsupported APIs in 1.5 besides these, so if you've used internal APIs in your apps, you need to update your apps to stop doing so. Even if they don't break on Android 1.5, there's a good chance they will on some later version. (There's some good news, though: because "flashlight" apps are so popular, we've added the "screenBrightness" field on the WindowManager.LayoutParams class just for that use case.)

Technique to Avoid, #2: Directly Manipulating Settings

Okay, strictly speaking this one isn't evil, since this is a change in behavior that we made to Android itself. But we made it because some developers were doing naughty things: a number of apps were changing system settings silently without even notifying the user. For instance, some apps turn on GPS without asking the user, and others might turn on data roaming.

As a result, applications can no longer directly manipulate the values of certain system Settings, even if they previously had permission to do so. For instance, apps can no longer directly turn on or off GPS. These apps won't crash, but the APIs in question now have no effect, and do nothing. Instead, apps will need to issue an Intent to launch the appropriate Settings configuration screen, so that the user can change these settings manually. For details, see the android.provider.Settings.Secure class, which you can find in the 1.5_pre SDK documentation (and later). Note that only Settings that were moved to the Settings.Secure class are affected. Other, less sensitive, settings will continue to have the same behavior as in Android 1.1.

Technique to Avoid, #3: Going Overboard with Layouts

Due to changes in the View rendering infrastructure, unreasonably deep (more than 10 or so) or broad (more than 30 total) View hierarchies in layouts are now likely to cause crashes. This was always a risk for excessively complex layouts, but you can think of Android 1.5 as being better than 1.1 at exposing this problem. Most developers won't need to worry about this, but if your app has very complicated layouts, you'll need to put it on a diet. You can simplify your layouts using the more advanced layout classes like FrameLayout and TableLayout.

Technique to Avoid, #4: Bad Hardware Assumptions

Android 1.5 includes support for soft keyboards, and there will soon be many devices that run Android but do not have physical keyboards. If your application assumes the presence of a physical keyboard (such as if you have created a custom View that sinks keypress events) you should make sure it degrades gracefully on devices that only have soft keyboards. For more information on this, keep on eye on this blog as we'll be posting more detailed information about handling the new soft keyboards.

Technique to Avoid, #5: Incautious Rotations

Devices running Android 1.5 and later can automatically rotate the screen, depending on how the user orients the device. Some 1.5 devices will do this by default, and on all others it can be turned on by the user. This can sometimes result in unpredictable behavior from applications that do their own reorientations (whether using the accelerometer, or something else.) This often happens when applications assume that the screen can only rotate if the physical keyboard is exposed; if the device lacks a physical keyboard, these apps do not expect to be reoriented, which is a coding error. Developers should be sure that their applications can gracefully handle being reoriented at any time.

Also, apps that use the accelerometer directly to reorient themselves sometimes compete with the system doing the same thing, with odd results. And finally, some apps that use the accelerometer to detect things like shaking motions and that don't lock their orientation to portrait or landscape, often end up flipping back and forth between orientations. This can be irritating to the user. (You can lock your app's orientation to portrait or landscape using the 'android:screenOrientation' attribute in your AndroidManifest.xml.)

Have any of your apps used one of these dubious techniques? If so, break out your IDE, duct tape, and spackle, and patch 'em up. I'm pretty excited by the new features in the Android 1.5 SDK, and I look forward to seeing your apps on my own 1.5-equipped phone -- but I can't, if they won't run! Fortunately, the fixes for these are pretty simple, and you can start fixing all of the above even with the 1.1_r1 SDK release.

By the way, if you'd like to fully immerse yourself in Android 1.5, join us at Google I/O! It's my pleasure to shamelessly plug an event that's shaping up to be the Android developer event of the year. We've added two more sessions—one on multimedia jujitsu, and a particularly interesting session on the Eyes-Free Android project—with even more yet to come. I thought Google I/O was a pretty killer event last year, and this year's looking even better, especially in terms of Android content.

I hope to meet many of you there, but either way, Happy Coding!

Creating an Input Method

To create an input method (IME) for entering text into text fields and other Views, you need to extend android.inputmethodservice.InputMethodService. This API provides much of the basic implementation for an input method, in terms of managing the state and visibility of the input method and communicating with the currently visible activity.

A good starting point would be the SoftKeyboard sample code provided as part of the SDK. Modify this code to start building your own input method.

An input method is packaged like any other application or service. In the AndroidManifest.xml file, you declare the input method as a service, with the appropriate intent filter and any associated meta data:

<manifest xmlns:android="http://schemas.android.com/apk/res/android"
package="com.example.fastinput">

<application android:label="@string/app_label">

<!-- Declares the input method service -->
<service android:name="FastInputIME"
android:label="@string/fast_input_label"
android:permission="android.permission.BIND_INPUT_METHOD">
<intent-filter>
<action android:name="android.view.InputMethod" />
</intent-filter>
<meta-data android:name="android.view.im" android:resource="@xml/method" />
</service>

<!-- Optional activities. A good idea to have some user settings. -->
<activity android:name="FastInputIMESettings" android:label="@string/fast_input_settings">
<intent-filter>
<action android:name="android.intent.action.MAIN"/>
</intent-filter>
</activity>
</application>
</manifest>

If your input method allows the user to tweak some settings, you should provide a settings activity that can be launched from the Settings application. This is optional and you may choose to provide all user settings directly in your IME's UI.

The typical life-cycle of an InputMethodService looks like this:

Visual Elements

There are 2 main visual elements for an input method—the input view and the candidates view. You don't have to follow this style though, if one of them is not relevant to your input method experience.

Input View

This is where the user can input text either in the form of keypresses, handwriting or other gestures. When the input method is displayed for the first time, InputMethodService.onCreateInputView() will be called. Create and return the view hierarchy that you would like to display in the input method window.

Candidates View

This is where potential word corrections or completions are presented to the user for selection. Again, this may or may not be relevant to your input method and you can return null from calls to InputMethodService.onCreateCandidatesView(), which is the default behavior.

Designing for the different Input Types

An application's text fields can have different input types specified on them, such as free form text, numeric, URL, email address and search. When you implement a new input method, you need to be aware of the different input types. Input methods are not automatically switched for different input types and so you need to support all types in your IME. However, the IME is not responsible for validating the input sent to the application. That's the responsibility of the application.

For example, the LatinIME provided with the Android platform provides different layouts for text and phone number entry:

InputMethodService.onStartInputView() is called with an EditorInfo object that contains details about the input type and other attributes of the application's text field.

(EditorInfo.inputType & EditorInfo.TYPE_CLASS_MASK) can be one of many different values, including:

  • TYPE_CLASS_NUMBER
  • TYPE_CLASS_DATETIME
  • TYPE_CLASS_PHONE
  • TYPE_CLASS_TEXT

See android.text.InputType for more details.

EditorInfo.inputType can contain other masked bits that indicate the class variation and other flags. For example, TYPE_TEXT_VARIATION_PASSWORD or TYPE_TEXT_VARIATION_URI or TYPE_TEXT_FLAG_AUTO_COMPLETE.

Password fields

Pay specific attention when sending text to password fields. Make sure that the password is not visible within your UI - in neither the input view nor the candidates view. And do not save the password anywhere without explicitly informing the user.

Landscape vs. portrait

The UI needs to be able to scale between landscape and portrait orientations. In non-fullscreen IME mode, leave sufficient space for the application to show the text field and any associated context. Preferably, no more than half the screen should be occupied by the IME. In fullscreen IME mode this is not an issue.

Sending text to the application

There are two ways to send text to the application. You can either send individual key events or you can edit the text around the cursor in the application's text field.

To send a key event, you can simply construct KeyEvent objects and call InputConnection.sendKeyEvent(). Here are some examples:

InputConnection ic = getCurrentInputConnection();
long eventTime = SystemClock.uptimeMillis();
ic.sendKeyEvent(new KeyEvent(eventTime, eventTime,
KeyEvent.ACTION_DOWN, keyEventCode, 0, 0, 0, 0,
KeyEvent.FLAG_SOFT_KEYBOARD|KeyEvent.FLAG_KEEP_TOUCH_MODE));
ic.sendKeyEvent(new KeyEvent(SystemClock.uptimeMillis(), eventTime,
KeyEvent.ACTION_UP, keyEventCode, 0, 0, 0, 0,
KeyEvent.FLAG_SOFT_KEYBOARD|KeyEvent.FLAG_KEEP_TOUCH_MODE));

Or use the convenience method:

InputMethodService.sendDownUpKeyEvents(keyEventCode);

Note: It is recommended to use the above method for certain fields such as phone number fields because of filters that may be applied to the text after each key press. Return key and delete key should also be sent as raw key events for certain input types, as applications may be watching for specific key events in order to perform an action.

When editing text in a text field, some of the more useful methods on android.view.inputmethod.InputConnection are:

  • getTextBeforeCursor()
  • getTextAfterCursor()
  • deleteSurroundingText()
  • commitText()

For example, let's say the text "Fell" is to the left of the cursor. And you want to replace it with "Hello!":

InputConnection ic = getCurrentInputConnection();
ic.deleteSurroundingText(4, 0);
ic.commitText("Hello", 1);
ic.commitText("!", 1);

Composing text before committing

If your input method does some kind of text prediction or requires multiple steps to compose a word or glyph, you can show the progress in the text field until the user commits the word and then you can replace the partial composition with the completed text. The text that is being composed will be highlighted in the text field in some fashion, such as an underline.

InputConnection ic = getCurrentInputConnection();
ic.setComposingText("Composi", 1);
...
ic.setComposingText("Composin", 1);
...
ic.commitText("Composing ", 1);

Intercepting hard key events

Even though the input method window doesn't have explicit focus, it receives hard key events first and can choose to consume them or forward them along to the application. For instance, you may want to consume the directional keys to navigate within your UI for candidate selection during composition. Or you may want to trap the back key to dismiss any popups originating from the input method window. To intercept hard keys, override InputMethodService.onKeyDown() and InputMethodService.onKeyUp(). Remember to call super.onKey* if you don't want to consume a certain key yourself.

Other considerations

  • Provide a way for the user to easily bring up any associated settings directly from the input method UI
  • Provide a way for the user to switch to a different input method (multiple input methods may be installed) directly from the input method UI.
  • Bring up the UI quickly - preload or lazy-load any large resources so that the user sees the input method quickly on tapping on a text field. And cache any resources and views for subsequent invocations of the input method.
  • On the flip side, any large memory allocations should be released soon after the input method window is hidden so that applications can have sufficient memory to run. Consider using a delayed message to release resources if the input method is in a hidden state for a few seconds.
  • Make sure that most common characters can be entered using the input method, as users may use punctuation in passwords or user names and they shouldn't be stuck in a situation where they can't enter a certain character in order to gain access into a password-locked device.

Samples

For a real world example, with support for multiple input types and text prediction, see the LatinIME source code. The Android 1.5 SDK also includes a SoftKeyboard sample as well.


Learn about Android 1.5 and more at Google I/O. Members of the Android team will be there to give a series of in-depth technical sessions and to field your toughest questions.

Updating Applications for On-screen Input Methods

One of the major new features we are introducing in Android 1.5 is our Input Method Framework (IMF), which allows developers on-screen input methods such as software keyboards. This article will provide an overview of what Android input method editors (IMEs) are, and what an application developer needs to do to work well with them. The IMF allows for a new class of Android devices, such as those without a hardware keyboard, so it is important that your application work well with it to provide the users of such devices a great experience.

What is an input method?

The Android IMF is designed to support a variety of IMEs, including soft keyboard, hand-writing recognizers, and hard keyboard translators. Our focus, however, will be on soft keyboards, since this is the kind of input method that is currently part of the platform.

A user will usually access the current IME by tapping on a text view to edit, as shown here in the home screen:

The soft keyboard is positioned at the bottom of the screen over the application's window. To organize the available space between the application and IME, we use a few approaches; the one shown here is called pan and scan, and simply involves scrolling the application window around so that the currently focused view is visible. This is the default mode, since it is the safest for existing applications.

Most often the preferred screen layout is a resize, where the application's window is resized to be entirely visible. An example is shown here, when composing an e-mail message:

The size of the application window is changed so that none of it is hidden by the IME, allowing full access to both the application and IME. This of course only works for applications that have a resizeable area that can be reduced to make enough space, but the vertical space in this mode is actually no less than what is available in landscape orientation, so very often an application can already accommodate it.

The final major mode is fullscreen or extract mode. This is used when the IME is too large to reasonably share space with the underlying application. With the standard IMEs, you will only encounter this situation when the screen is in a landscape orientation, although other IMEs are free to use it whenever they desire. In this case the application window is left as-is, and the IME simply displays fullscreen on top of it, as shown here:

Because the IME is covering the application, it has its own editing area, which shows the text actually contained in the application. There are also some limited opportunities the application has to customize parts of the IME (the "done" button at the top and enter key label at the bottom) to improve the user experience.

Basic XML attributes for controlling IMEs

There are a number of things the system does to try to help existing applications work with IMEs as well as possible, such as:

  • Use pan and scan mode by default, unless it can reasonably guess that resize mode will work by the existence of lists, scroll views, etc.
  • Analyze the various existing TextView attributes to guess at the kind of content (numbers, plain text, etc) to help the soft keyboard display an appropriate key layout.
  • Assign a few default actions to the fullscreen IME, such as "next field" and "done".

There are also some simple things you can do in your application that will often greatly improve its user experience. Note that, except where explicitly mentioned, all of the things suggested here will not tie your application to Android 1.5 -- it will still work on older releases, which will simply ignore these new options.

Specifying each EditText control's input type

The most important thing for an application to do is use the new android:inputType attribute on each EditText, which provides much richer information about the text content. This attribute actually replaces many existing attributes (android:password, android:singleLine, android:numeric, android:phoneNumber, android:capitalize, android:autoText, android:editable); if you specify both, Cupcake devices will use the new android:inputType attribute and ignore the others.

The input type attribute has three pieces:

  • The class is the overall interpretation of characters. The currently supported classes are text (plain text), number (decimal number), phone (phone number), and datetime (a date or time).
  • The variation is a further refinement on the class. In the attribute you will normally specify the class and variant together, with the class as a prefix. For example, textEmailAddress is a text field where the user will enter something that is an e-mail address (foo@bar.com) so the key layout will have an '@' character in easy access, and numberSigned is a numeric field with a sign. If only the class is specified, then you get the default/generic variant.
  • Additional flags can be specified that supply further refinement. These flags are specific to a class. For example, some flags for the text class are textCapSentences, textAutoCorrect, and textMultiline.

As an example, here is the new EditText for the IM application's message text view:

    <EditText android:id="@+id/edtInput"
android:layout_width="0dip"
android:layout_height="wrap_content"
android:layout_weight="1"
android:inputType="textShortMessage|textAutoCorrect|textCapSentences|textMultiLine"
android:imeOptions="actionSend|flagNoEnterAction"
android:maxLines="4"
android:maxLength="2000"
android:hint="@string/compose_hint"/>

A full description of all of the input types can be found in the documentation. It is important to make use of the correct input types that are available, so that the soft keyboard can use the optimal keyboard layout for the text the user will be entering.

Enabling resize mode and other window features

The next most important thing for you to do is specify the overall behavior of your window in relation to the input method. The most visible aspect of this is controlling resize vs. pan and scan mode, but there are other things you can do as well to improve your user experience.

You will usually control this behavior through the android:windowSoftInputMode attribute on each <activity> definition in your AndroidManifest.xml. Like the input type, there are a couple different pieces of data that can be specified here by combining them together:

  • The window adjustment mode is specified with either adjustResize or adjustPan. It is highly recommended that you always specify one or the other.
  • You can further control whether the IME will be shown automatically when your activity is displayed and other situations where the user moves to it. The system won't automatically show an IME by default, but in some cases it can be convenient for the user if an application enables this behavior. You can request this with stateVisible. There are also a number of other state options for finer-grained control that you can find in the documentation.

A typical example of this field can be see in the edit contact activity, which ensures it is resized and automatically displays the IME for the user:

    <activity name="EditContactActivity"
android:windowSoftInputMode="stateVisible|adjustResize">
...
</activity>

For non-activity windows, there is a new Window.setSoftInputMode() method that can be used to control their behavior. Note that calling this API will make your application incompatible with previous Android platforms.

Controlling the action buttons

The final customization we will look at is the "action" buttons in the IME. There are currently two types of actions:

  • The enter key on a soft keyboard is typically bound to an action when not operating on a mult-line edit text. For example, on the G1 pressing the hard enter key will typically move to the next field or the application will intercept it to execute an action; with a soft keyboard, this overloading of the enter key remains, since the enter button just sends an enter key event.
  • When in fullscreen mode, an IME may also put an additional action button to the right of the text being edited, giving the user quick access to a common application operation.

These options are controlled with the android:imeOptions attribute on TextView. The value you supply here can be any combination of:

  • One of the pre-defined action constants (actionGo, actionSearch, actionSend, actionNext, actionDone). If none of these are specified, the system will infer either actionNext or actionDone depending on whether there is a focusable field after this one; you can explicitly force no action with actionNone.
  • The flagNoEnterAction option tells the IME that the action should not be available on the enter key, even if the text itself is not multi-line. This avoids having unrecoverable actions like (send) that can be accidentally touched by the user while typing.
  • The flagNoAccessoryAction removes the action button from the text area, leaving more room for text.
  • The flagNoExtractUi completely removes the text area, allowing the application to be seen behind it.

The previous IM application message view also provides an example of an interesting use of imeOptions, to specify the send action but not let it be shown on the enter key:

 android:imeOptions="actionSend|flagNoEnterAction"

APIs for controlling IMEs

For more advanced control over the IME, there are a variety of new APIs you can use. Unless special care is taken (such as by using reflection), using these APIs will cause your application to be incompatible with previous versions of Android, and you should make sure you specify android:minSdkVersion="3" in your manifest.

The primary API is the new android.view.inputmethod.InputMethodManager class, which you can retrieve with Context.getSystemService(). It allows you to interact with the global input method state, such as explicitly hiding or showing the current IME's input area.

There are also new window flags controlling input method interaction, which you can control through the existing Window.addFlags() method and new Window.setSoftInputMode() method. The PopupWindow class has grown corresponding methods to control these options on its window. One thing in particular to be aware of is the new WindowManager.LayoutParams.FLAG_ALT_FOCUSABLE_IM constant, which is used to control whether a window is on top of or behind the current IME.

Most of the interaction between an active IME and application is done through the android.view.inputmethod.InputConnection class. This is the API an application implement, which an IME calls to perform the appropriate edit operations on the application. You won't normally need to worry about this, since TextView provides its own implementation for itself.

There are also a handful of new View APIs, the most important of these being onCreateInputConnection() which creates a new InputConnection for an IME (and fills in an android.view.inputmethod.EditorInfo structure with your input type, IME options, and other data); again, most developers won't need to worry about this, since TextView takes care of it for you.


Learn about Android 1.5 and more at Google I/O. Members of the Android team will be there to give a series of in-depth technical sessions and to field your toughest questions.