To support a wide range of writing styles and poorly formed cursive script, the Tablet PC input panel uses a time-delay neural network working with a lexicon. An improved recognizer supports both personalization and error reporting.

In “Handwriting Recognition: Tablet PC Text Input,” Jay Pittman declares that pen-based computing is here now and describes how a large neural network trained on a very large training set containing ink samples from thousands of people with a diverse range of writing styles is being used to support poorly formed cursive script, providing amazing recognition accuracy, especially once users acclimate to writing with a plastic pen on a plastic screen.

In "Facilitating Pedagogical Practices through a Large-Scale Tablet PC Deployment," Joe Tront describes the multifaceted, collaborative approach that faculty and students at the Virginia Tech College of Engineering are using. They have begun exploration of the use of Tablet PCs in engineering and computer science courses, working with an implementation process that includes computer acquisition, faculty training, infrastructure modifications, and multiple learning assessments. Preliminary results have been quite positive.

Cn specific situations, machine-recognized handwriting serves as a useful alternative for entering small volumes of text. However, practical handwriting recognition isn’t easy. Not only must handwriting-recognition systems handle many different shapes and styles for each letter, but humans also commonly produce sloppy script to increase throughput speed. To support a wide range of writing styles and poorly formed cursive script, the Tablet PC input panel uses a timedelay neural network working with a lexicon. High-end versions of Microsoft’s Vista now include Tablet PC software, with an improved handwriting recognizer that supports both personalization and error reporting.