Ionic Launch Event. September 28th 2021.
For more details on the history of Cordova and more details on how it works, please see here.
The open source space is filled with new projects that build on top of older projects, making tangible improvements that can’t be done without radically changing the original product. This is what Capacitor would have required. The Ionic team felt like this wasn’t possible with Cordova for technical and political reasons. Whether that is right or wrong, that is the conclusion the team came to. That said, Ionic still uses Cordova heavily and will continue to invest in the platform for a long time to come.
On the plus side, Ionic now controls almost all of its stack. When you build an Ionic app and use Capacitor, we control the native runtime layer, the UI controls, and the “framework” used to build the controls (Stencil). The only part we don’t control is the frontend framework you use on top (Angular, React, Vue, or nothing). This is significant: If there’s an issue in any part of the stack that we control, we can fix it right away. Capacitor has already proven to be a worthwhile investment - it’s enabling us to build a stronger Ionic and focus on what we do uniquely well.
In spirit, Capacitor and Cordova are very similar. Both manage a Web View and provide a structured way of exposing native functionality to your web code. However, Capacitor has a few key differences that require web developers, previously used to Cordova’s approach, to change app development workflows.
Capacitor considers each platform project a source asset instead of a build time asset. That means you’ll check your Xcode and Android Studio projects into source control, as well as use those IDEs when necessary for platform-specific configuration and running/testing.
This change in approach has a few implications. First, Capacitor does not use
config.xml or a similar custom configuration for platform settings. Instead, configuration changes are made by editing the appropriate platform-specific configuration files directly, such as
AndroidManifest.xml for Android and
Info.plist for Xcode. Capacitor does have some
high level configuration options that are set in
capacitor.config.json. These generally don’t modify native functionality, but control Capacitor’s tooling.
Additionally, Capacitor does not “run on device” or emulate through the command line. Instead, such operations occur through the platform-specific IDE, which provides a faster, more typical experience that follows the standards of app development for that platform. For example, running iOS apps from the command line is not officially supported by Apple anyway, so Xcode is preferred.
While these changes may be concerning to long-time Cordova users, there are worthwhile benefits:
config.xmlis error prone and a constant moving target. Becoming more comfortable with platform-specific tooling makes troubleshooting issues that much easier.
Capacitor manages plugins in a different way than Cordova. First, Capacitor does not copy plugin source code to your app before building. Instead, all plugins are built as Frameworks (on iOS) and Libraries (on Android) and installed using the leading dependency management tool for each platform (CocoaPods and Gradle/Maven, respectively). Additionally, Capacitor does not modify native source code, so any necessary native project settings must be added manually (for example, permissions in
AndroidManifest.xml). We think this approach is less error-prone and makes it easier for developers to find help in the community for each specific platform.
deviceready event. As soon as your app code loads, you can start calling plugin methods.
exec() in Cordova, the plugin will reference the Capacitor plugin through
Finally, Capacitor has implications for plugin authors. On iOS, Swift 4 is officially supported and even
preferred for building plugins (Objective-C is also supported). Plugins no longer export a
Plugin.xml file; Capacitor provides a few simple macros on iOS and Annotations on Android for adding metadata to your plugin source code that Capacitor reads at runtime.
Capacitor, unlike Cordova, does not use a global CLI. Instead, the Capacitor “CLI” is installed locally into each project as an npm script. This makes it easier to manage versions of Capacitor across many different apps.
Thus, instead of running
capacitor directly from the command line, Capacitor is invoked by calling
npx cap in the directory of your app.