UNIPEN data collection

The following image shows how a word-oriented data-collection program may look like. Note that you are not going to collect completely natural writing behaviour with this example. The layout and interface are designed to extract isolated words of reasonable quality from the writer. The same words may be written quite differently - by the very same writer - in the context of a sentence or paragraph!

Note the progress bar at the top of the LCD screen of the pen computer. It enables the users to see how far they are in the process. The target word (here: "last") is presented at the top. The rectangle with the dotted line is for entering the word in any style (print, mixed, cursive). If the word which just has been written is considered illegible by the writer him/herself, tapping on the Cancel button allows for a rewrite. To call for the next word, the Next button must be tapped. In this screen dump, the buttons currently are disabled (green) because no ink has been produced, as yet. The program of this particular example has been written in VB with a dedicated DLL in C, which uses Windows 3.1 for Pen Computing.

The dotted line is carefully placed. Never use two dotted lines for trying to 'box' cursive handwriting. Writers will then produce multi-stable character sizes, i.e., sometimes fitting the ascenders and descenders within the two lines, at other times using the lines for fitting the small lowercase characters and leaving the ascenders and descenders to stick out. The single dotted line is a good solution for eliciting a fairly horizontal orientation of the words.

Below is the same screen dump with real-size dimensions:

(How this picture was created? You do a print screen on the notepad with you collection tool running, then load the clipboard contents into Paintbrush, and write with the pen on the screen dump .BMP image to annotate it. Isn't pen computing great?)

Back to the UNIPEN home page at NICI

schomaker@computer.org