Conventional text fields

The most useful as well as the most common form input element is the text entry field. A text entry field appears in the browser window as an empty box on one line and accepts a single line of user input that becomes the value of the element when the user submits the form to the server. To create a text entry field inside a form in your HTML document, set the type of the <input> form element to text. Include a name attribute as well; it's required.

What constitutes a line of text differs among the various browsers. Fortunately, HTML gives us a way, with size and maxlength attributes, to dictate the width, in characters, of the text input display box, and how many total characters to accept from the user, respectively. The value for either attribute is an integer equal to the maximum number of characters you'll allow the user to see and type in the field. If maxlength exceeds size, then text scrolls back and forth within the text entry box. If maxlength is smaller than size, there will be extra blank space in the text entry box to make up the difference between the two attributes.

The default value for size is dependent upon the browser; the default value for maxlength is unlimited. We recommend you set them yourself. Adjust the size attribute so that the text entry box does not extend beyond the right margin of a typical browser window (about 60 characters with a very short prompt). Set maxlength to a reasonable number of characters; for example, 2 for state abbreviations, 12 for phone numbers, and so on.

A text entry field is usually blank at first until the user types something into it. You may, however, specify an initial default value for the field with the value attribute. The user may modify the default, of course. If the user presses a form's reset button, the value of the field is reset to this default value. [the section called "Reset buttons"]

These are all valid text entry fields:

<input type=text name=comments>
<input type=text name=zipcode size=10 maxlength=10>
<input type=text name=address size=30 maxlength=256>
<input type=text name=rate size=3 maxlength=3 value="100">

The first example creates a text entry field set to the browser's default width and maximum length. As we argue above, this is not a good idea because defaults vary widely among browsers, and your form layout is sure to look bad with some of them. Rather, fix the width and maximum number of acceptable input characters as we do in the second example: it lets the user type in up to ten characters inside an input box ten characters wide. Its value will be sent to the server with the name "zipcode" when the user submits the form.

The third example field tells the browser to display a 30-character-wide text input box into which the user may type up to 256 characters. The browser automatically scrolls text inside the input box to expose the extra characters.

The last text input field is three characters wide, lets the user type in only three characters, and sets its initial value to 100.

Notice in the second and fourth example fields, it is implied that certain kinds of data are to be entered by the user--a postal code or a numeric rate, respectively. Except for limiting how many, HTML provides no way for you to dictate what characters may be typed into a text input field. For instance, in the last example field, the user may type "ABC" even though you intend it to be a number less than 1,000. Your server-side application must trap erroneous or mistaken input, as well as check for incomplete forms, and send the appropriate error message to the user when things aren't right. That can be a tedious process, so we emphasize again, provide clear and precise instructions and prompts. Make sure your forms tell users what kinds of input you expect from them, thereby reducing the number of mistakes they may make when filling it out.

Masked text fields

Like the Lone Ranger, the mask is on the good guys in a masked text field. It behaves just like a conventional text field in a form, except that the user-typed characters don't appear onscreen. Rather, the browser obscures the characters in a masked text to keep such things as passwords and other sensitive codes from prying eyes.

To create a masked text field, set the value of the type attribute to password. All other attributes and semantics of the conventional text field apply to the masked field. Hence, you must provide a name, and you may (we recommend it) specify a size and maxlength for the field, as well as an initial value.

Don't be misled: a masked text field is not all that secure. The typed-in value is only obscured onscreen; the browser transmits it unencrypted when the form is submitted to the server. So, while prying eyes may not see them onscreen, devious bad guys may steal the information electronically.

file selection fields

File selection fields

As its name implies, the field lets users select a file stored on their computer and send it to the server when they submit the form. Although part of the HTML 3.2 standard and documented by Microsoft as being supported, file-selection fields do not work in Internet Explorer 3.0. They do work in Netscape.

The browser presents the file-selection form field to the user like other text fields. With Netscape, it's accompanied by a button labeled "Browse" to its right. Users either type the pathname directly as text into the field or, with the Browse option, select the name of a locally stored file from a system-specific dialog box.

Create a file-selection field in a form by setting the value of the type attribute to file. Like other text fields, the size and maxlength of a file-selection field should be set to appropriate values, with the browser creating a field 20 characters wide, if not otherwise directed. Since file and directory names differ widely among systems, it makes no sense to provide a default value for this field. As such, the value attribute should not be used with this kind of text field.

The Browse button associated with Netscape's file-selection field opens a platform-specific file-selection dialog that allows users to select a value for the field. In this case, the entire pathname of the selected file is placed into the field, even if the length of that pathname exceeds the field's specified maxlength.

Use the accept attribute to constrain the types of files that the browser lets the user select. It's value is a comma-separated list of MIME encodings; users can only select files whose type matches one of those in the list. For example, to restrict the selection to images, you might add accept="image/*" to the file selection <input> tag.

Unlike other form input elements, the file-selection field only works correctly with a specific form data encoding and transmission method. If you include one or more file-selection fields in your form, you must set the enctype attribute of the <form> tag to multipart/form-data and the <form> tag's method attribute to post. Otherwise, the file-selection field behaves like a regular text field, transmitting its value (that is, the file's pathname) to the server instead of the contents of the file itself.

This is all easier than it may sound. For example, here is a form that collects a person's name and favorite file:

<form enctype="multipart/form-data" method=post
    action="cgi-bin/save_file">
Your name: <input type=text size=20 name=the_name>
<p>
Your favorite file: <input type=file size=20 name=fav_file>
</form>

The data transmitted from the browser to the server for this example form has two parts. The first contains the value for the name field, and the second contains the name and contents of the specified file:

-----------------------------6099238414674
Content-Disposition: form-data; name="the_name"
  
One line of text field contents
-----------------------------6099238414674
Content-Disposition: form-data; name="fav_file"; filename="abc"
  
First line of file
...
Last line of file
-----------------------------6099238414674--

The browsers don't check that a valid file has been specified by the user. If no file is specified, the filename portion of the Content-Disposition header will be empty. If the file doesn't exist, its name appears in the filename subheader, but there will be no Content-Type header or subsequent lines of file content. Valid files may contain nonprintable or binary data; there is no way to restrict user-selectable file types. In light of these potential problems, the form-processing application on the server should be robust enough to handle missing files, erroneous files, extremely large files, and files with unusual or unexpected formats.

Submission buttons

The submit button (<input type=submit>) does what its name implies, setting in motion the form's submission to the server from the browser. You may have more than one submit button in a form. You may also include name and value attributes with the submit type of input form button.

With the simplest submit button (that without a name or value attribute), the browser displays a small rectangle or oval with the default label "Submit" (see Figure 10.1). Otherwise, the browser will label the button with the text you include with the tag's value attribute. If you provide a name attribute, the value attribute for the submit button will be added to the parameter list the browser sends along to the server. That's good, because it gives you a way to identify which button in a form was pressed, letting you process any one of several different forms with a single form-processing application.

The following are all valid submission buttons:

<input type=submit>
<input type=submit value="Order Kumquats">
<input type=submit value="Ship Overnight" name="ship_style">

The first one is also the simplest: the browser displays a button, labeled "Submit," which activates the form-processing sequence when clicked by the user. It does not add an element to the parameter list that the browser passes to the form-processing server and application.

The second example button has the value attribute that makes the displayed button label "Order Kumquats," but like the first example, does not include the button's value in the form's parameter list.

The last example sets the button label and makes it part of the form's parameter list. When clicked by the user, that last example of the submission button adds the parameter ship_style="Ship Overnight" to the form's parameter list.

reset buttons

Reset buttons

The reset type of form <input> button is nearly self-explanatory: it lets the user reset--erase or set to some default value--all elements in the form. Unlike the other buttons, a reset button does not initiate form processing. Instead, the browser does the work of resetting the form elements. The server never knows (or cares, for that matter) if or when the user might have pressed a reset button.

By default, the browser displays a reset button with the label "Reset." You can change that by specifying a value attribute with your own button label.

Here are two sample reset buttons:

<input type=reset>
<input type=reset value="Use Defaults">

The first one creates a reset button labeled "Reset"; the browser labels the second example reset button with "Use Defaults." They both initiate the same reset response in the browser.

Custom image buttons

The image type of <input> form element creates a custom button that is a "clickable" image. It's a special button made out of your specified image that, when clicked by the user, tells the browser to submit the form to the server, and includes the x,y coordinates of the mouse pointer in the form's parameter list, much like the mouse-sensitive image maps we discuss in Chapter 7, Links and Webs. Image buttons require a src attribute with the URL of the image file, and you can include a name attribute. You may also use align and border (Netscape only) attributes to control image alignment within the current line of text and the width of the frame that Netscape places around the image, respectively, much like the align and border attributes for the <img> tag (Internet Explorer doesn't border form <input> images). Interestingly, none of the JavaScript event handlers work with this type of form element, despite what the manufacturer's documentation says--not even onClick, which you might expect, since the image acts like the submit input type.

Here are a couple of valid image buttons:

<input type=image src="pics/map.gif" name=map>
<input type=image src="pics/xmap.gif" align=top name=map>

The browser displays the designated image within the form's content flow. The second button's image will be aligned with the top of the adjacent text, as specified by the align attribute. Some browsers (Netscape, for instance) also add a border, as it does when an image is part of an anchor (<a> tag), to signal that the image is a form button.

When the user clicks the image, the browser sends the horizontal offset, in pixels, of the mouse from the left edge of the image and the vertical offset from the top edge of the image to the server. These values are assigned the name of the image as specified with the name attribute, followed by ".x" and ".y," respectively. Thus, if someone clicked the image specified above, the browser would send parameters named map.x and map.y to the server.

Image buttons behave much like mouse-sensitive image maps, and, like the programs that process image maps, your form-processing application may use the x,y mouse-pointer parameters to choose a special course of action. You should use an image button when you need additional form information to process the user's request. If an image map of links is all you need, use a mouse-sensitive image map. Mouse-sensitive images also have the added benefit of providing server-side support for automatic detection of shape selection within the image, letting you deal with the image as a selectable collection of shapes. Buttons with images require you to write code that determines where the user clicked on the image and how this position can be translated to an appropriate action by the server.

Regular buttons

Using the <input type=button> tag, you can create a button that can be clicked by the user but that does not submit or reset the form. The value attribute can be used to set the label on the button; the name attribute, if specified, will cause the supplied value to be passed to the form processing script.

You might wonder what value such buttons provide. Little or none, unless you supply the onClick attribute, along with a snippet of JavaScript to be executed when the user clicks the button. Thus empowered, regular buttons can be used to validate form contents, update fields, manipulate the document, and initiate all sorts of client-side activity. [the section called "JavaScript Event Handlers"]

Multiple buttons in a single form

You can have several buttons of the same or different types in a single form. Even simple forms have both reset and submit buttons, for example. To distinguish between them, make sure each has a different value attribute, which the browser uses for the button label. Depending on the way you program the form-processing application, you might also make the name of each button different, but it is usually easier to name all similarly acting buttons the same and let the button handling subroutine sort them out by value. For instance:

<input type=submit name=action value="Add">
<input type=submit name=action value="Delete">
<input type=submit name=action value="Change">
<input type=submit name=action value="Cancel">

When the user selects one of these example buttons, a form parameter named action will be sent to the server. The value of this parameter will be one of the button names. The server-side application gets the value and behaves accordingly.

Since an image button doesn't have a value attribute, the only way to distinguish between several image buttons on a single form is to ensure they all have different names.