Let’s look at the “First Name” field first. What are the validations that we need to apply? The data on this form feeds into our billing system, so the customer’s name is very important—the field must be required. Are there any other validations? Not only do we want to get the data, but also we’d like it to be in a consistent format. Possibly the data should be stored in all capital letters. Or maybe we want to ensure that the data is not in all capitals. Let’s choose the latter—but we’ll still want to make sure that at least the first letter is capitalized. As in many of the issues related to validation, things are more complicated then they might first appear. For example, are we allowing enough room to enter long names? Will single-word names such as “Bono” be allowed? For our purposes, we’ll keep it simple.

We turn on validation by using special attribute values in the HTML markup for these fields. The following code will add validation to the fields.

    <label for="firstName">First Name: </label>
    <input type="text" id="firstName" name="firstName"

        dojoType="dijit.form.ValidationTextBox"
        required="true"
        propercase="true"
        promptMessage="Enter first name."
        invalidMessage="First name is required."
        trim="true"

    /><br>

The code is formatted to be more readable by using line breaks. To summarize what has happened: All we’ve done is add some new attributes to the <input> tag for the field. Each of the new attributes affects the validation in some way.

Notice the following line of code from the preceding example:

       dojoType="dijit.form.ValidationTextBox"

This attribute is not a standard HTML <input> tag attribute. Depending on which editor you are using to modify the file, it may even be highlighted as an error. The dojoType attribute is only meaningful to the Dojo parser, which was referenced in step 1. Remember the code we needed to include the parser? It is shown here:

       dojo.require("dojo.parser");

The parser reads through the HTML and looks for any tag that contains dojoType as one of its attributes. Then the magic happens. The parser replaces the element with the Dojo widget specified by dojoType. In this case, the widget dijit.form.ValidationTextBox is substituted for the Document Object Model (DOM) element created from the <input> tag.

How does Dojo know what to replace the tag with? That is determined by the specific widget. Each widget behaves a little differently. HTML markup and JavaScript code is associated with the widget in its definition, and that is how Dojo knows what to replace the original element with—which brings us to the missing piece of the puzzle. We need to tell Dojo to include the code for the widget by specifying the widget in JavaScript. To do that, we include the following JavaScript code after the link to Dojo and after the reference to the Dojo parser.

    dojo.require("dijit.form.ValidationTextBox");

Notice that the name of the widget specified as the value for the dojoType attribute is the same as the argument for the dojo.require call. This is the linkage that allows Dojo to associate the HTML markup with the JavaScript code for that widget.

To emphasize this process, let’s review the HTML markup specified in the original page and then compare it to the HTML markup after the parser runs. To see the original markup, we merely have to view the source of the file form.html. Seeing the new markup is a bit harder. The browser converts the original HTML into a DOM tree representing the various tags. The Dojo parser modifies the DOM elements using JavaScript, but the original source for the page is untouched. We need some tool that will convert the DOM (the browser’s internal representation of the page) back into HTML for our review. The Firefox browser provides a DOM Inspector to do just that. An excellent add-on to Firefox, called Firebug, also allows the DOM to be inspected. Firebug also provides a number of excellent tools for developing web pages such as its DOM inspection capabilities we can use to inspect the DOM after the Dojo parser has run—so we can see exactly what it does. But before we see how the DOM changes, let’s first review the original <input> tag for the first name field.

    <input
        type="text"
        id="firstName"
        size="20"
        dojoType="dijit.form.ValidationTextBox"
        required="true"
        propercase="true"
        promptMessage="Enter first name."
        invalidMessage="First name is required."
        trim="true"
    />

The code has been reformatted to make it more readable by adding some line breaks. The attributes from dojoType through trim are not valid HTML attributes. They are meaningful only to the Dojo parser and drive some features of the Dojo widget they pertain to. Now let’s see what the HTML looks like after the parser runs.

<input
  type="text"
  tabindex="0"
  maxlength="999999"
  size="20"
  class="dijitInputField dijitInputFieldValidationError dijitFormWidget"
  name="firstName"
  id="firstName"
  autocomplete="off"
  style=""
  value=""
  disabled="false"

  widgetid="firstName"
  dojoattachevent="onfocus,onkeyup,onkeypress:_onKeyPress"
  dojoattachpoint="textbox,focusNode"
  invalid="true"
  valuenow=""

/>

The preceding code has also been reformatted for readability, adding line breaks and changing the order of the attributes a little. Notice that a number of valid HTML attributes have been added to the <input> DOM element such as tabindex, class, autocomplete, and disabled. And additionally, a number of Dojo-only attributes have been added such as widgetid, dojoattachevent, dojoattachpoint, invalid, and valuenow. We look at these in more detail in Part II, “Dojo Widgets,” but for now it’s enough just to point out that the parser is rewriting our HTML. The parser is doing even more work that we can see here. It is associating various event handler functions to events that might occur on this DOM element. For instance, when the user enters or changes the value in the field, Dojo functions get called, which perform validation. And Dojo even creates objects that correspond to the HTML tags. We can’t tell that this is happening just from seeing the HTML markup, but behind the scenes, that is exactly what Dojo is doing.

Let’s review the other special Dojo attributes. Each Dojo widget has a set of properties that control its behavior. These properties are set by various Dojo widget attribute values.

Now let’s run the page and see how it behaves. Because this is the first field on the page, the field gets focus, and the cursor immediately is placed on the input area for the “First Name” field.

Notice that we get a message box that says “Enter first name.” Dojo calls this a Tool Tip, and it has dynamic behavior. It is only displayed when the field has focus (the cursor is in the field), and once the field loses focus, the message disappears. The message appears on top of any visible element below it, so there is no need to leave room for it when designing your page.

Try entering different values in the field and then press <tab> to leave the field. For example, enter “joe” and watch it be transformed into “Joe” with leading and trailing spaces removed and the first letter of the name capitalized.

The last name field has the same validations as the first name field does. There is nothing extra to do for this field and nothing new to learn. Just replace the <input> tag for Last Name with the following code.

    <input type="text" id="lastName" name="lastName"
        dojoType="dijit.form.ValidationTextBox"
        required="true"
        propercase="true"
        promptMessage="Enter last name."
        invalidMessage="Last name is required."
        trim="true"
    />

We are going to allow the user to manage his or her own account information in our application. To provide some security we need the user to make up a user name that he or she can use later to sign on to the system. This field will be required, and we’d like it to always be entered in lowercase. To validate this field, we’ll use the same Dojo widget that we’ve already used, dijit.form.ValidationTextBox, but we’ll use a new attribute called lowercase to force the transformation of the entered data into all lowercase letters.

There are some additional validations we’d like to do on this field. For instance, is this user name already assigned to someone else? We could check the server for existing values. However, because this validation requires interaction with the server, we’ll save it for step 3 of the tutorial and focus on only the client-side validation right now.

The following HTML markup is needed to enable validation for this field.

    <input type="text" id="userName" name="userName"
        dojoType="dijit.form.ValidationTextBox"
        required="true"
        promptMessage="Enter user name."
        trim="true"
        lowercase="true"
    />

We need to communicate with our customers so we’ll get their email addresses. This will be a required field. We’ll also make it all lowercase for consistency. In addition, we’d like to make sure that the value entered in this field is also in the correct format for an email address. There is no way to know if it is a working email until we actually try to send something to it, but at least we can make sure that it contains a “@” character and appears to reference a valid domain.

How can we specify the desired format? By using a specialized pattern matching language known as regular expressions, we can specify a pattern of characters to check the value against. We need to build a regular expression to validate for email addresses. At this point in our discussions, let’s not go on a long detour to discuss the building of these expressions.

The following is regular expression that can be used to validate most formats of email addresses—most because it is surprisingly difficult to validate for all possible email addresses. This is because of some of the unusual variations such as domains longer than four characters such as “.museum” or addresses consisting of a sub-domain. But the following regular expression will work for most.

[[A-Z0-9._%+-]+@[A-Z0-9.-]+.[A-Z]{2,4}]+

The ValidationTextBox contains a special property for validating against regular expressions. The attribute to use is regExp—just specify the regular expression as its value. Replace the <input> tag for email with the following code in “form.html” to specify validation for the email address field.

    <input type="text" id="email" name="email" size="30"
        dojoType="dijit.form.ValidationTextBox"
        required="true"
        regExp="[a-zA-Z0-9._%-]+@[a-zA-Z0-9.-]+.[a-zA-Z]{2,4}"
        promptMessage="Enter email address."
        invalidMessage="Invalid Email Address."
        trim="true"
    />

Validating email addresses is a really interesting subject. There are quite a few variants to the simple [email protected] format that we often see. For a really thorough discussion of email, you should review the RFC rules. The following link will get you to the Wikipedia page that describes email, from which you can link to the official RFC documents: http://en.wikipedia.org/wiki/E-mail_address.

The address field will contain the first line of the user’s mailing address. We’ll make it required. We will use the ValidationTextBox, and we have seen all of the attributes already. Replace the <input> tag for address with the following code.

    <input type="text" id="address" name="address" size="30"
        dojoType="dijit.form.ValidationTextBox"
        required="true"
        promptMessage="Enter address."
        invalidMessage="Address is required."
        trim="true"
    />

There are many additional validations that can be performed on address data, the most important being to ensure that the address is an actual address. Standard abbreviations such as “St” for “Street” could also be allowed. These additional validations could be done by a number of web services available from the U.S. Postal Service, but that is really outside the scope of this tutorial.

The city field will contain the value for the city in the user’s mailing address. We’ll make it required. We will use the ValidationTextBox. Replace the <input> tag for address with the following code.

    <input type="text" id="city" name="city" size="30"
        dojoType="dijit.form.ValidationTextBox"
        required="true"
        promptMessage="Enter city."
        invalidMessage="City is required."
        trim="true"
    />

The zip code field is part of the mailing address and is required. There are some additional validations we can apply. Our hypothetical company is a U.S. corporation and only provides service to U.S. customers, so we’ll limit our address to valid U.S. addresses, which means that the zip code must be in one of two forms. Either it is a 5-digit number, or it is a 5-digit number followed by a dash and then followed by a 4-digit number. If we can come up with a regular expression to test for either format, then we’re golden!

Replace the <input> tag for zip code with the following to enable Dojo validation for this field.

    <input type="text" id="zipCode" name="address" size="30"
        dojoType="dijit.form.ValidationTextBox"
        trim="true"
        required="true"
        regExp="d{5}([-]d{4})?$"
        maxlength="10"
        promptMessage="Enter zip code."
        invalidMessage="Invalid zip code (NNNNN) or (NNNNN-NNNN)."
    />

An interesting feature of the preceding code is that we’ve got two overlapping validations. The maxlength attribute prevents the value from being over 10 digits, but so does that regular expression. What are the implications of this? One could argue that it is inefficient because both validations will be executed. But they each operate differently on the page, which might justify using both. If the user tries to enter a zip code that is 12 digits long, he will be notified as he tries to type the eleventh digit, rather than after typing all 12 digits and pressing tab to leave the field. By using both techniques, the error is detected sooner.

       dojoType="dijit.form.ValidationTextBox"

dojo.require("widget dijit.form.ValidationTextBox");

In this chapter we’ve focused on functionality that doesn’t require a call to the server. In the next chapter the server will play a role. We’ll make calls to the server using the XMLHttpRequest to get data and perform validations. Now that’s Ajax!