App Inventor

CHAPTER 14

Understanding an App’s Architecture

This chapter examines the structure of an app from a programmer’s perspective. It begins

with the traditional analogy that an app is like a recipe and then proceeds to reconcep-

tualize an app as a set of components that respond to events. The chapter also examines

how apps can ask questions, repeat, remember, and talk to the Web, all of which will be

described in more detail in later chapters.

Many people can tell you what an app is from a user’s perspective, but understand-

ing what it is from a programmer’s perspective is more complicated. Apps have an

internal structure that we must fully understand in order to create them effectively.

One way to describe an app’s internals is to break it into two parts, its components

and its behaviors . Roughly, these correspond to the two main windows you use in

App Inventor: you use the Component Designer to specify the objects (components)

of the app, and you use the Blocks Editor to program how the app responds to user

and external events (the app’s behavior).

Figure 14-1 provides an overview of this app architecture. In this chapter, we’ll explore

this architecture in detail.

220  Chapter 14: Understanding an App’s Architecture

App

Components

Visible

Button, Textbox,

Label, etc.

Non-visible

Texting,

LocationSensor,

Text To Speech

Button.Click,

Texting.MessageReceived

Event Response

Ball.MoveTo,

Texting.SendMessage,

set Label.Visible to

Variables

Behaviors:

Event Handlers

Procedures

Figure 14-1. The internal architecture of an App Inventor app

Components

There are two main types of components in an app: visible and non-visible. The app’s

visible components are the ones you can see when the app is launched—things like

buttons, text boxes, and labels. These are often referred to as the app’s user interface .

Non-visible components are those you can’t see, so they’re not part of the user in-

terface. Instead, they provide access to the built-in functionality of the device; for ex-

ample, the Texting component sends and processes SMS texts, the LocationSensor

component determines the device’s location, and the TextToSpeech component

talks. The non-visible components are the technology within the device—little

people that do jobs for your app.

Both visible and non-visible components are defined by a set of properties . Properties

are memory slots for storing information about the component. Visible components,

for instance, have properties like Width , Height , and Alignment , which together

define how the component looks. So a button that looks like the Submit button in

Figure 14-2 to the end user is defined in the Component Designer with a set of prop-

erties, including those shown in Table 14-1.

Table 14-1. Button properties

Width Height Alignment Text

50 30 center Submit

Figure 14-2. Submit button

Behavior  221

You can think of properties as something like the cells you see in a spreadsheet. You

modify them in the Component Designer to define the initial appearance of a com-

ponent. If you change the number in the Width slot from 50 to 70, the button will

appear wider, both in the Designer and in the app. Note that the end user of the app

doesn’t see the 70; he just sees the button’s width change.

Behavior

An app’s components are generally straightforward to understand: a text box is for

entering information, a button is for clicking, and so on. An app’s behavior, on the

other hand, is conceptually difficult and often complex. The behavior defines how

the app should respond to events, both user initiated (e.g., a button click) and exter-

nal (e.g., an SMS text arriving to the phone). The difficulty of specifying such interac-

tive behavior is why programming is so challenging.

Fortunately, App Inventor provides a visual “blocks” language perfectly suited for

specifying behaviors. This section provides a model for understanding it.

An App As a Recipe

Traditionally, software has often been compared to a recipe. Like a

recipe, a traditional app follows a linear sequence of instructions,

such as those shown in Figure 14-3, that the computer (chef) should

perform.

A typical app might start a bank transaction (A), perform some

computations and modify a customer’s account (B), and then print

out the new balance on the screen (C).

An App As a Set of Event Handlers

However, most apps today, whether they’re for mobile phones,

the Web, or desktop computers, don’t fit the recipe paradigm anymore. They don’t

perform a bunch of instructions in a predetermined order; instead, they react to

events—most commonly, events initiated by the app’s end user. For example, if the

user clicks a button, the app responds by performing some operation (e.g., sending a

text message). For touchscreen phones and devices, the act of dragging your finger

across the screen is another event. The app might respond to that event by drawing

a line from the point of your original touch to the point where you lifted your finger.

These types of apps are better conceptualized as a set of components that respond

to events. The apps do include “recipes”—sequences of instructions—but each

recipe is only performed in response to some event, as shown in Figure 14-4.

A

B

C

Figure 14-3.

Traditional

software follows a

linear sequence of

instructions

222  Chapter 14: Understanding an App’s Architecture

A

B

C

Event1

D

E

Event2

Figure 14-4. An app as multiple recipes hooked to events

So, as events occur, the app reacts by calling a sequence of functions . Functions are

things you can do to or with a component—operations like sending an SMS text, or

property-changing operations such as changing the text in a label of the user inter-

face. To call a function means to invoke it, to make it happen. We call an event and the

set of functions performed in response to it an event handler .

Many events are initiated by

...