The following is the first few sections of a chapter from The Busy Coder's Guide to Android Development, plus headings for the remaining major sections, to give you an idea about the content of the chapter.
Android is an ever-moving target, averaging about 2.5 API level increments per year. The Android Developer site maintains a chart and table showing the most recent breakdown of OS versions making requests of the Play Store.
Most devices tend to be clustered around 1-3 minor releases. However, these are never the most recent release, which takes time to percolate through the device manufacturers and carriers and onto devices, whether those are new sales or upgrades to existing devices.
Some developers panic when they realize this.
Panic is understandable, if not necessary. This is a well-understood problem, that occurs frequently within software development — ask any Windows developer who had to simultaneously support everything from Windows 98 to Windows XP, or Windows XP through Windows 8.1. Moreover, there are many things in Android designed to make this problem as small as possible. What you need are the strategies and tactics to make it all work out.
Android itself tries very hard to maintain backwards compatibility. While each new Android release adds many classes and methods, relatively few are marked as deprecated, and almost none are outright eliminated. And, in Android, “deprecated” means “there’s probably a better solution for what you are trying to accomplish, though we will maintain this option for you as long as we can”.
Despite this, many developers aim purely for the lowest common denominator. Aiming to support older releases is noble. Ignoring what has happened since those releases is stupid, if you are trying to distribute your app to the public via the Play Store or similar mass-distribution means.
Why? You want your app to be distinctive, not decomposing.
For example, as we saw in the chapter on the action bar, adding one line to
the manifest (
android:targetSdkVersion="11") gives you the action bar, the
holographic widget set (e.g.,
Theme.Holo), the new style of options menu,
and so on. Those dead-set on avoiding things newer than Android 2.1 would not
use this attribute. As a result, on Android 3.0+ devices, their apps will
tend to look old. Some will not, due to other techniques they are employing
(e.g., running games in a full-screen mode), but many will.
You might think that this would not matter. After all, how many people in 2011 were even using Android 3.x? 5%?
However, those in position to trumpet your application — Android enthusiast bloggers chief among them — will tend to run newer equipment. Their opinion matters, if you are trying to have their opinion sway others relative to your app. Hence, if you look out-of-touch to them, they may be less inclined to provide glowing recommendations of your app to their readers.
Besides, not everything added to newer versions of Android is pure “eye candy”. It is entirely possible that features in the newer Android releases might help make your app stand out from the competition, whether it is making greater use of NFC or offering tighter integration to the stock Calendar application or whatever. By taking an “old features only” approach, you leave off these areas for improvement.
And, to top it off, the world moves faster than you think. It takes about a year for a release to go from release to majority status (or be already on the downslope towards oblivion, passed over by something newer still). You need to be careful that the decisions you make today do not doom you tomorrow. If you focus on “old features only”, how much rework will it take you to catch up in six months, or a year?
Hence, this book advocates an approach that differs from that taken by many: aim high. Decide what features you want to use, whether those features are from older releases or the latest-and-greatest release. Then, write your app using those features, and take steps to ensure that everything still works reasonably well (if not as full-featured) on older devices. This too is a well-trodden path, used by Web developers for ages (e.g., support sexy stuff in Firefox and Safari, while still gracefully degrading for IE6). And the techniques that those Web developers use have their analogous techniques within the Android world.
One thing to bear in mind is that the OS distribution chart and table shown above is based on devices contacting the Play Store. Hence, this is only directly relevant if you are actually distributing through the Play Store.
If you are distributing through the Amazon AppStore, or to device-specific outlets (e.g., BlackBerry World), you will need to take into account what sorts of devices are using those means of distribution.
If you are specifically targeting certain non-Play Store devices, like the Kindle Fire, you will need to take into account what versions of Android they run.
If you are building an app to be distributed by a device manufacturer on a specific device, you need to know what Android version will (initially) be on that device and focus on it.
If you are distributing your app to employees of a firm, members of an organization, or the like, you need to determine if there is some specific subset of devices that they use, and aim accordingly. For example, some enterprises might distribute Android devices to their employees, in which case apps for that enterprise should run on those devices, not necessarily others.
The preview of this section was eaten by a grue.
The preview of this section apparently resembled a Pokémon.
The preview of this section was whisked away by a shark-infested tornado.
The preview of this section was abducted by space aliens.
The preview of this section was stepped on by Godzilla.
The preview of this section is being chased by zombies.