Newton OS 1.x: Handwriting Recognition Setup and Tips

I would like to improve handwriting recognition on my Newton MessagePad. What steps do I follow to setup recognition? What steps should I follow to improve recognition reliability?
The Newton OS will learn to recognize your handwriting gradually, but you likely will have to teach it your handwriting style both letter by letter and word by word.

These steps will help your Newton recognize your handwriting:
Be sure that the text recognizer button is turned on. The text recognizer button is at the bottom-left side of the screen, marked with an A. When it is on, it appears as a white A on a black background.

Handwriting Style
-----------------
To teach the Newton OS to better understand your handwriting style, follow these steps:

1) Tap the Extras drawer icon at the bottom of the screen, then tap Prefs.
2) Tap Handwriting Style in the Preferences Menu.
3) Tap the button that best describes your handwriting under "My handwriting style is". You may find that "Mixed Cursive & Printed works best, unless you exclusively use all-printed or all cursive writing.
4) Adjust the other preferences so they reflect your writing style. Drag the diamonds with the stylus to move them.
5) Make sure that the box at the bottom marked "Configure for guest user" is unselected.

Individual Letters
------------------
To teach the Newton OS how you write individual letters, follow these steps:

1) Tap the Extras drawer at the bottom of the screen, then tap Prefs.
2) Tap Letter Styles in the Preferences Menu. You will see a dialog box which displays a particular letter, number or symbol written in several different styles. You can select which character to view at the bottom of the dialog box.
3) Select each character, then tap each of the samples and carefully watch how the Newton OS draws the letter. Indicate which style or styles you use most often, and which you use rarely. When you finish, tap the next letter, number or symbol you wish to train the Newton OS to learn to recognize, and continue.

Entire Words
------------
To teach the Newton OS how you write entire words, follow these steps:

1) Tap the Extras drawer at the bottom of the screen, then tap Prefs.
2) Tap Handwriting Practice in the Preferences Menu. In this dialog box, a prompt will ask you to write a particular word and try to recognize it. The more words you practice, the more the Newton OS learns.

Reviewing your Personal Word List
---------------------------------
To review the words in your personal word list, follow these steps:

1) Tap the Extras drawer at the bottom of the screen, then tap Prefs.
2) Tap Recognizers in the Preferences Menu, then click on "Review my word list". You may choose to delete words you rarely or never use, so Newton has fewer words to compare your writing to before it interprets. Add often-used words which don't appear in a dictionary, such as proper names, abbreviations and slang terms.

Recognizers
-----------
There are other choices in the Recognizers Menu that will aid handwriting recognition.

For the best speed and accuracy, select the box next to "words in word lists" and deselect "words not in word lists." This limits the number of words Newton considers when recognizing text.

Also, deselect as many of the other options such as "names," "telephone numbers," "dates," and "times of day" as possible. This will improve word recognition, particularly when writing in the Notepad and not using numbers.

If you need to write a number in the Notepad and these options are deactivated, tap the screen where you want the number to appear, then tap the keyboard icon at the bottom of the screen, and type the number in.


The following handwriting tips will help the Newton OS recognize your handwriting:

Note: After restoring your data from your Newton Messagepad to your MessagePad 100 or MessagePad 110, go to the preferences section and reset your letter styles training.

You can do this by choosing Preferences in the Extras drawer, choosing Letter Styles, and tapping on "Return to original settings." Your recognition should be better than it was on the old unit even with the 'old' training data.
Published Date: Feb 18, 2012