Interfacing with Personal Information Management (PIM) Applications in BlackBerry

0
36
16 min read

(For more resources on BlackBerry, see here.)

The BlackBerry handhelds come pre-loaded with many great programs to help make a person more productive. While messages may be the most common reason a person will purchase a BlackBerry, the other Personal Information Management (PIM) applications often quickly become essential as well. Not only can you interface with these applications by adding or editing content in them, you can also listen for events that allow you to react to things that happen. Some of the applications even allow you to add menu items and other “active content” to them. That’s a lot to talk about, so we’ll just focus on some of the most common tasks to get you started in this article.

As a developer, you cannot ignore the other applications on the handheld. The more integrated an application can be with these standard applications, the better the user experience will generally be. Our TipCalc application is very specialized, and one of the few that works well without integrating with other applications. More often than not though, any applications that you create will benefit from some level of integration.

Introducing PIM

The first area that we will take a look at is the Personal Information Management, or PIM applications and data. PIM applications are a rather generic name for a group of tools that manage your personal information, especially as it relates to your handheld. This could be stretched to include a lot of things, but it generally means your messages, contacts, and scheduling information that help you to manage your life. In BlackBerry terms it means the Messages, Address Book, Calendar, Tasks, and Notes.

Access to the PIM data in the BlackBerry SDK is provided through the JSR-75 specification, which is a Java standard. Like many of the Java standards in the BlackBerry SDK, there are also BlackBerry-specific extensions available that expand the basic classes with new BlackBerry-specific functionality.

Like many of the other standards we find in Java, JSR-75 implements a factory pattern where one class, in this case the PIM class, is used to create objects for the other more specific types of PIM data. The PIM class can basically do only one thing and that is to retrieve a PIMList object that contains a bunch of specialized PIMItem objects.

Why is all so generic?

All of these PIM classes may seem very generic and you would be absolutely correct. They are generic and they are supposed to be that way. PIM data is a very generic concept so the implementation is very generic as well. Also, because it is a Java standard, it needs to be flexible enough to accommodate any system that it might be implemented on.

A perfect example of this kind of flexibility is the BlackBerry PIN field. The BlackBerry PIN is an entry in your address book and therefore, it should be included in the PIM data that you get. However, a PIN is a BlackBerry-specific concept and no other device out there will use it. You can’t really expect the Java standard to include specialized fields for every possible piece of data that some device may have or want to include. The answer to this is to present PIM data in a key-value pairing so that it is flexible enough to handle every possible scenario.

A key-value pairing is a somewhat technical term to describe the pattern for storing values based on a static key. Or, more simply, if you know the proper key you can access the value. The flexible part is that the PIM object storing all of the values does not need to know about each specific value or provide any special mechanism for accessing each specific value. All access is done through generic methods, which also require the key.

The difficulty in using this kind of approach is that the keys must be common knowledge. In addition, simple numeric keys do not support self-documenting code or even easily readable code. Keys that are string values offer a lot of advantages—in that the keys are much more readable, but the possibility for mistakes is very great because you don’t have the compiler to help ensure that only correct keys are used.

To help solve these issues there are derived classes for each type of PIM item. While you can do nearly everything by using the generic classes that a derived class would offer, I wouldn’t recommend it. These classes are here to make your code easier to write and read, and should be used.

PIMLists

As we said early on, the PIM class is used primarily to retrieve a PIMList object for a specific kind of PIM item, that is, address book contact, calendar event, and so on. For each of these types, there is also a specialized class that you can use instead of the generic PIMList class. Classes such as ContactList, EventList, and ToDoList offer a specialized version of the more generic PIMList class. These specialized classes are also part of the Java standard and should be preferred because they offer a few more methods which are specific to that kind of data.

There are BlackBerry-specific versions of these classes as well. Therefore, the BlackBerryContactList class is the BlackBerry-specific version of the ContactList, which is in turn a specialized version of PIMList for contact data. Generally speaking, you will want to use the BlackBerry-specific versions of PIMList classes when making your applications.

PIMItems

A PIMItem is the generic representation for any piece of PIM data. Just like there are specific versions of the PIMList, there are also specific versions of the PIMItem class for each kind of PIMItem. Contact, Event, and ToDo are all specific versions of a PIMItem for that kind of PIM data. As you might expect, there are BlackBerry-specific versions as well. BlackBerryContact, BlackBerryEvent, and BlackBerryToDo all exist to extend and further specialize the standard Java classes.

These specialized classes aren’t as specialized or easy to use as one might expect though. Providing a method called getFirstName might be really useful, but unfortunately, you will find nothing of the sort. The specialized classes offer few methods for accessing data. Instead, they provide static values for the keys used to set and retrieve data from the PIMItem class. Remember, earlier we noted that one drawback to using this kind of key-value pairing was that keys were sometimes not clear and that you could not expect help from the compiler. By providing each key value in the specialized class, both of these goals are accomplished. The name of the key value now provides a readable name and the compiler will flag an error if there is a typo or problem with the constant value being used.

Another aspect of PIMItem is that each value that an item has a specific type associated with it as well. Some of these are obvious, such as the start date of an event using a Date type. Some of them, such as the Name field of a Contact that requires an array, are not. Some fields can be given a subtype as well, such as the Phone field. With the subtype you can specify what kind of phone number it is: home, work, or mobile. Furthermore, some of the fields can have multiple occurrences while others cannot. A good example of this is the Email field in a Contact. A contact is allowed to have up to three e-mail addresses, but there is no subtype associated with them like there is with phone numbers. The bottom line to all this is that the PIM items have a defined structure to them and they won’t allow just any value to be put into a field. The documentation plays a big role here in understanding this because there are no field-specific methods to provide some additional assistance in the proper way to access each field.

Laying the ground work

Still, this is all rather abstract because you haven’t seen any real code samples yet, so let’s get into some code! For this article, you will build an application that someone will use to request some time off from their manager. This is definitely one of those applications that just can’t be done without interfacing with other applications on the handheld! To make getting started a little easier we will take the starter TimeOff application from the code bundle and add to it throughout this article.

The first task to undertake is one to help make testing and debugging easier. Remember, you will be working on the simulator, which is essentially a brand new device and which can be often reset. That means you don’t have any of your contacts there! You will need some contacts later, so to get started let’s add a menu item to the application that will create a few contacts that you can later use to test with.

Time for action – creating test contacts

  1. Modify the _AddTestAddressesAction menu item in the TimeOff project to look like the following completed code.
  2. protected MenuItem _AddTestAddressesAction = new MenuItem(
    "Add Test Data", 800000, 50)
    {
    public void run()
    {
    PIM pimInstance = PIM.getInstance();
    try
    {
    // TODO: Create test contacts
    BlackBerryContactList contacts =
    (BlackBerryContactList)pimInstance.openPIMList(
    PIM.CONTACT_LIST, PIM.READ_WRITE);
    BlackBerryContact newContact1 =
    (BlackBerryContact)contacts.createContact();
    BlackBerryContact newContact2 =
    (BlackBerryContact)contacts.createContact();

    String[] names = new
    String[contacts.stringArraySize(BlackBerryContact.NAME)];
    names[BlackBerryContact.NAME_FAMILY] = "Smith";
    names[BlackBerryContact.NAME_GIVEN] = "John";
    if (contacts.isSupportedArrayElement(Contact.NAME,
    Contact.NAME_SUFFIX))
    {
    names[BlackBerryContact.NAME_SUFFIX]="Jr";
    }
    newContact1.addStringArray(
    BlackBerryContact.NAME,
    BlackBerryContact.ATTR_NONE,
    names);
    names[Contact.NAME_FAMILY] = "Doe";
    names[Contact.NAME_GIVEN] = "John";
    if (contacts.isSupportedArrayElement(Contact.NAME,
    Contact.NAME_PREFIX))
    {
    names[Contact.NAME_PREFIX] = "Dr.";
    }
    newContact2.addStringArray(Contact.NAME, Contact.ATTR_NONE,
    names);
    //TODO: Add Phone numbers
    //TODO: Add Email Addresses
    //TODO: Add Addresses
    newContact1.commit();
    newContact2.commit();
    }
    catch (PIMException e)
    {
    // TODO Auto-generated catch block
    e.printStackTrace();
    }
    }
    };

  3. Then add this line to the constructor to make the menu item available when you run the application.

    this.addMenuItem(_AddTestAddressesAction);

What just happened?

This is the first of several baby steps as you work towards the goal of creating some test contacts in the address book in the simulator. As the address book in the simulator doesn’t have any entries to begin with, and can be erased frequently, doing this provides you with a way to quickly and easily create or recreate the contacts you will use later on for testing other parts of this application. It also happens to be a great way to demonstrate how to add contacts.

This first baby step does only a few things. First, it gets a PIMList object for contacts and then creates two new contacts. After this it sets the name of each one and finally commits the records into the address book. These steps make sense at a high level, but let’s take a look at the details.

The first step is to get an instance to the PIM object, which is done through a static method in the PIM class called getInstance.

PIM pimInstance = PIM.getInstance();

Once you have an instance of PIM, the next step is to get a list of contact items using the openPIMList method on the instance you just retrieved. This same method is used to get a list of any kind of PIM item so you must specify the type of data to get as one of the parameters. The PIM class offers constant values for every kind of PIM item, so in this case, use the constant PIM.CONTACT_LIST. As you plan to add new contacts, the next parameter needs to be the constant PIM.READ_WRITE so that you have write permissions. It’s always good practice to request the minimum amount of permissions that you need, so if your application doesn’t change or add data to the list you should simply use the PIM.READ permission.

As we touched on earlier, this method returns a generic PIMList type so you also have to cast it to the appropriate specialized type. If a list type of CONTACT_LIST has been specified, you can cast the resulting PIMList to either of the available specialized classes—ContactList or BlackBerryContactList. As long as your application is a BlackBerry-specific application, there is no good reason to use the less specialized class of ContactList. Instead, you should always use BlackBerryContactList.

BlackBerryContactList contacts =
(BlackBerryContactList)pimInstance.openPIMList
(PIM.CONTACT_LIST, PIM.READ_WRITE);

The next step is to create a couple of new contacts that you will start to populate. This is done through the createContact method available on the ContactList class. Again, you need to cast the resulting objects to the proper type. The createContact method returns a contact, but again you’ve chosen to use the more specialized version of BlackBerryContact instead. Because this is all being executed on a BlackBerry handheld, you can always cast a contact to a BlackBerryContact safely. The same is true for each of the Java specialized classes and their corresponding BlackBerry specialized class.

BlackBerryContact newContact1 = (BlackBerryContact) contacts.
createContact();
BlackBerryContact newContact1 = (BlackBerryContact) contacts.
createContact();

The next segment of code sets the name attribute of the newly created contacts. Notice that this is done through an array of String objects instead of individual methods. This isn’t something that is done to be more efficient, it is done this way because it must be; there is no other way.

We mentioned before about each field in a PIMItem having a type associated with it. Most of the field types are basic String or Date type fields, but NAME is more complicated than most of the other fields. The NAME field is defined as a StringArray because there are many parts to a name and you want to be able to set each part separately. There aren’t very many fields of this type used, but this is probably one of the most important. You can only set the NAME field as a whole unit, so if only one part of the name needs to be changed the entire name field must be replaced.

To work with the name, you must first create a string array of the proper size. There is no constant value for this as it may vary with the SDK version. Instead, you must first get the size by using the stringArraySize method on the ContactList and then construct a new array by using the returned value.

String[] names = new String[contacts.stringArraySize(BlackBerryContact.NAME)];

Once you have an array of the proper size each part of the name is set by indexing the array by using the NAME constant from the Contact class.

names[BlackBerryContact.NAME_FAMILY] = "Smith";
names[BlackBerryContact.NAME_GIVEN] = "John";

In this example, you also want to add another name part but are not sure whether the field is supported in this system. Not all fields are supported and not all of the name subfields are supported either. You can test to see whether a field or a subfield is supported by using the isSupportedField or isSupportedArrayElement methods in the ContactList class. In this case, you test to see if the suffix is supported, and then set the suffix if so.

if (contacts.isSupportedArrayElement(Contact.NAME,
Contact.NAME_SUFFIX))
{
names[BlackBerryContact.NAME_SUFFIX]="Jr";
}

This step is very important if you want to use the same code for multiple platforms. Each system can support the fields it chooses. In this case, the suffix is NOT supported and if you were to step through this code in the debugger, you would see that the code to set the suffix is skipped over. Later on, when you test this application, you will also see that the suffix was not added to the contact.

Other platforms may implement it differently. You could just assume each of the name subfields are supported and set the field without testing to see if it is supported. In the BlackBerry SDK, unsupported fields are just quietly ignored. This can lead to confusion wondering why a field doesn’t appear in the Address Book application, but it won’t cause an error.

The next step is to actually add the NAME field to the contact. Up until this time you’ve simply been building an array in memory with all of the proper values.

newContact1.addStringArray(
BlackBerryContact.NAME,
BlackBerryContact.ATTR_NONE,
names);

Notice that the method addStringArray doesn’t give any indication about what field is being added, but only what type of data is being added. All of the PIMItem methods are like this. Remember, this class is designed to be generic. The first parameter is the field indicator, which is one of the many constants that are defined in the Contact class. In this case, we use the BlackBerryContact class. Because BlackBerryContact derived from Contact, all of the constant values are accessible. The BlackBerryContact class does define some constants that are BlackBerry-specific, such as PIN. For this field you must reference the constant value from BlackBerryContact because the Java standard Contact class does not define it. Partly for this reason, I suggest always referencing constant values from BlackBerryContact because all of the constant values will be available through this class.

The method addStringArray was chosen because that is the type of data that you are adding. The NAME field is defined as a string array and so you must use the addStringArray method because it corresponds to the data type of the field.

Once you finish with the first contact, the code starts building the NAME string array to add a second contact. For demonstration sake, all of the constant values that are referenced are done so using the Contact class instead of the BlackBerryContact class.

names[Contact.NAME_FAMILY] = "Doe";
names[Contact.NAME_GIVEN] = "John";
if (contacts.isSupportedArrayElement(Contact.NAME,
Contact.NAME_PREFIX))
{
names[Contact.NAME_PREFIX] = "Dr.";
}
newContact2.addStringArray(Contact.NAME, Contact.ATTR_NONE, names);

Also, notice that the second contact applies a prefix to the name and tests to see if it is supported in the same way as you did for the suffix when adding the previous contact. However, the prefix is supported and if you were to step through this method in the debugger, you would see that the prefix is being set properly.

The last step you have to do is to commit the data that has been added to the contact.

newContact1.commit();
newContact2.commit();

Simply creating a new contact is not enough; you must commit the changes in it by using the commit method. Creating a contact and then never committing it will not have any effect on the Address Book application. It simply won’t be shown in the list. That’s the whole point of this exercise, so you have to make sure and commit the changes once they are all done.

At this point, if you were to run the application and select the menu, you would see two new contacts added to the Address Book application in the simulator. They would show up as Dr. John Doe and John Smith. There would be only names with these contacts because that is all that you’ve added so far.

In the example code that you just stepped through there was one mistake that could have proven to be very serious. Did you catch it? You are reusing the names array to set the names of both contacts. This is actually risky, but it happens to work out in this case. If the SUFFIX field had been supported then your Dr. John Doe would have actually been Dr. John Doe Jr. because the array was not reset before it was used again. If you had changed the order around, John Smith would have been Dr. John Smith. This might have lead to a bug that could have been tough to track down, so keep it in mind.

LEAVE A REPLY

Please enter your comment!
Please enter your name here