TEXT, NATIVE LANGUAGE SUPPORT, AND TIME MEDIA

How input methods work

An input method is used as part of the typing process provided by the typing configuration mechanism. You--or in later releases, an end user--select a particular input method for inclusion in a typing configuration. The typing configuration controls the process for converting input text into ideograms. Typically, the input process is:

The user types in text phonetically using a different script--for example, Roman.
One or more keyboard transliterators transcribe the input into phonetic symbols, or components, for the target language--for example, Kana for Japanese or Hangul for Korean.
The input method converts the phonetic transcriptions into the appropriate ideograms.

The entire process occurs inline in an active area within the document. The active area should be visually differentiated so the user can see which text is being processed. The active area can contain both raw, unconverted text, exactly as the user entered it, and converted text. The user can then adjust the converted text and, when it is correct, confirm it. Confirmed text becomes part of the document and is no longer part of the active area. When the user has confirmed all of the input text, the active area collapses completely. The text editing application defines the appearance of the active area and the interface for adjusting and confirming input text. For an example, see "Entering Japanese text" on page 188; this section describes the user interface for entering text with the Writing Heads input method.

Input method operations

An input method can support a variety of operations that allow the user to adjust text before confirming it. You add these operations using the class TInputMethodTextPresentation, described in "Supporting input methods in a text editor" on page 191. The following are some operations you might support when building an interface for using an input method:

Selecting a phrase--Before adjusting a phrase, the user needs to select it. You can implement a selection mechanism based on either mouse interaction or keyboard commands, such as the arrow keys. After the user selects a phrase, it should be visually differentiated from both the text in the body of the document and the other text in the current active area.
Changing a phrase boundary--The user may want to make a particular phrase longer or shorter by dragging the boundary with the mouse or using keyboard commands. For some input methods, this will cause the boundaries of subsequent phrases to change as well.
Choosing a homophone--To adjust the ideograms for a phrase, the user needs a mechanism for displaying and selecting alternate homophones. The user might display alternate homophones by double-clicking on the phrase or by using a sequence of keyboard commands. The user might select an alternate homophone by choosing the next closest homophone or by viewing a window with all homophones and selecting the correct one.
Confirming the conversion--The user must confirm that the conversion has been successfully completed--for example, with a command such as the Return or Enter key. After the text is confirmed, the active area should collapse and the insertion point caret should be placed after the newly input text.

Input method dictionaries

Input methods use dictionaries to perform the text conversions. These dictionaries contain entries that map components, or groups of phonetic characters, to ideographic characters. Each entry has associated grammatical information that helps the input method choose the most appropriate result for a given context.

An input method uses a single main dictionary that provides common information for a particular language. Typically, the main dictionary cannot be edited by the user. An input method can also use one or more user dictionaries. User dictionaries are designed to be editable, allowing a user to create a customized dictionary--for example, providing special terminology or giving the input method hints for selecting results most suited to the user's style or usage.

The Input Method framework is designed to allow input method dictionaries to work with more than one input method. This means users can buy specialized dictionaries--for example, for particular fields such as law or medicine--and plug them into their favorite input method. A future release of the CommonPoint application system will provide protocols to allow you to create input method dictionaries.

[Contents] [Previous] [Next]

Click the icon to mail questions or corrections about this material to Taligent personnel.

Generated with WebMaker