Hi … for what it;s worth … my thoughts …
I have used Processing/Android Mode for a few years now and whilst it makes it very easy (once you have a working setup!) to code and run an Android app, to do more advanced stuff you are better in Android Studio. It depends what you want to do. I see you are using with students … what age and what level and what are the learning goals? If you are aiming to teach Android development then Android Studio is best, but if it is beginner programming (especially youngsters) with the fun element of being able to have a phone app, then Processing with Android Mode is a great way in. Processing itself isn’t java, although the basic rules syntax etc is the same so it’s a good starting point. The power of Processing is that it is incredibly easy to have something working with very little effort, which is great for beginners. Having said that, you can do lot with processing, and develop skills/knowledge that readily transfers to 'proper’java. Android Mode is the same … it hides the tricky things that make learning coding difficult, but can, for example, import from the android api if the needed functionality isn’t available by default. This is quite a good way into exploring the android sdk.
Re creating new Activities in a Processing sketch … I don’t think you can do this. Android mode just provides a Fragment that you code within … basically allowing for the coding of of a single view (I think!). I have never found a way to add activities or fragments within Processing. If I’m wrong, then I would be keen to know how! To get round this, I use a switch variable (displayMode is what I call it) within draw() so that different pieces of code run in draw() for differing screens.
Books are hard to find!. There is a good one from a few years back that covered the Ketai contributed library, but I think Ketai is now a bit behind Processing, with a number of compatibility issues. It may be OK now but when it started not working for me after an Android Mode upgrade, I decided to dig into the android sdk directly. That’s the trouble with contributed libraries … great if they are continually updated etc but no good long term if they’re not. It’s also a good way into ‘proper’ android for your students if they have to access the sdk documentation!
Some things are easy to do … eg making a Button. You could go to the android sdk and learn how to do this with all the various setup requirements and listener callbacks etc but a very easy way is in Processing to define your own simple Button class. Give it methods to draw rectangle or display an image, and a method to return a boolean TRUE of it is clicked (or tapped). … called from mousePressed() or alternatively touchStarted() in Android Mode.
Enough for now …
Happy to help more if you have specific questions.