|
 |
TransNote transformation
By Victor Chase
| |
 Though the much-touted paperless society has not yet come to pass, IBM Research and IBM's Personal Computing Division (PCD) have co-developed a machine that integrates computers and paper in previously unimaginable harmony.
The new machine, introduced in February, opens like a book to reveal a ThinkPad® on one side and a seemingly conventional notepad on the other — all rolled into one small portfolio-sized package called TransNoteTM. As the user applies ink to paper, a digital facsimile of the handwriting created by tracking the movements of the pen is captured and uploaded to the ThinkPad in a proprietary IBM format, which can then be converted into popular formats such as JPEG, TIFF and PDF.
An overnight success years in the making
Like many of the innovative products that ultimately emerge from IBM Research laboratories, TransNote's technology was already in hand; the challenge was applying it. Though the product launched only 18 months after researchers began work on it, TransNote evolved from pure research that began years earlier.
In 1996, as Research completed a prototype of an electronic notepad, the makers of Cross pens came to IBM asking for just such a product. A marriage of convenience followed, and CrossPad® 1.0 was born shortly thereafter. It entered the world as an 8 1/2 X 11-inch digitizing surface that included memory, allowing the CrossPad to connect to a PC for downloading data. Cross sold about 100,000 of these tablets under a licensing agreement with IBM.
| | 
Randy Moulic with TransNote prototype
|
Meanwhile, Research's Mobile PC Platforms group was working closely with PCD, looking for ways to breathe new life into the ThinkPad. Linking the CrossPad with the ThinkPad made sense, and a joint Research/PCD prototyping effort commenced. To expedite product development, Randy Moulic, manager of client systems, merged Jayashree Subrahmonia's pen technologies group from Research and Dan Dumarot's mobile PC platforms group to work on the project.
Because the underlying technology of speech and handwriting recognition is similar, TransNote's ancestry goes back some 25 years, to the time that IBM first began its speech recognition efforts, notes Subrahmonia. It wasn't until 1991, however, that a concerted effort to develop Very Large Vocabulary Unconstrained (VLVU) handwriting recognition was launched, sowing the seeds for TransNote.
Therein lies much of the value of Research, according to Moulic. "The work that we do here daily tends to build the core basis of the technologies that can really provide new products," he said.
The VLVU effort involved capturing and recognizing natural print and cursive handwriting, and has led to a 30,000-word vocabulary that TransNote uses as a font of keywords for conversion to text.
The early work also led to a natural pen and paper interface for getting "ink" into a PC. It consists of a digitizing surface that serves as a base upon which a standard writing pad is placed, and includes a built-in antenna. A pen containing a transmitter sends signals to the base as the user writes on the pad, and the strokes are captured in digital form. The notepad portion of TransNote uses the same concept.
Research plays new roles
| |
"We are doing uncommon things because it's important for research to be relevant. The way TransNote appears, the way it is used, its design, even its weight and size were driven by customer input."
Mark Dean, VP & IBM Fellow, Systems Research
|
Researchers involved with TransNote contributed to its design by playing a significant part in customer focus groups held during the early days of the effort. Conducted in New York, Chicago, Tokyo and Paris, these sessions aimed to determine the viability of the concept and discern which features potential buyers found most important.
"We worked closely with PCD to organize and design these focus groups," said Mark Dean, Research vice president. Research also provided mock-ups of the TransNote concept, which were shown to potential customers. Researchers then observed the response and were heavily involved in the analysis of the feedback in shaping iterations of the prototype.
As a result of this process, "The way TransNote appears, the way it is used, its design, even its weight and size were driven by customer input," said Dean.
According to Scott Lekuch, a researcher involved in the customer interface, it was an iterative process involving a good deal of rapid prototyping. "Usually in Research," he said, "we tend to do exclusive testing, not towards a product, but towards a more conceptual mode. Here there were real-world constraints that we couldn't ignore. Things like how do you fit the hardware into an acceptable form factor, and what is that form factor? With pure research you can ignore some of those constraints and take it as an assumption that you will be able to do it eventually. In working with PCD on a real product these were not concerns we could dismiss."
Product Development: Research gets involved
While the creation of TransNote was itself no small technological feat, the process of its development set a noteworthy precedent for researchers' involvement in product development.
"It's very rare that you see a whole Research organization involved to the extent that we were in doing the TransNote product work," said Dan Dumarot, manager of mobile PC platforms.
There were some struggles in the development of TransNote. But, just as the device brings paper and computer together, so, too, did this project lead to a blending of Research's more academic culture with the product orientation of PCD, Research's internal customer for this project. According to Moulic, the small group of researchers wasn't accustomed to spending half of their time on phone calls and in meetings deciding what they wanted to do. Working in close proximity to product development was a cultural change for the Research staff.
Subrahmonia recalls that before TransNote, "I had no clue about all of the different cycles involved in getting a product out the door." But things are different now. "It was a huge learning experience, and it helped us a great deal," she said.
Importantly, researchers are continuing their cooperation with customers, in part through a software development kit enabling users to adapt the system to their own applications.
In an example of this innovation a large insurance company recently sent six developers to Yorktown to work with Research on creating software for their forms they use for writing farm policies.
Agents must first walk the farm noting information such as the number of pigs and sheep, and the size of chicken coops. Lekuch helped the company develop software aimed at enabling agents to fill out their forms on TransNote and then upload the handwritten data into the ThinkPad for transmission to the home office. The system is currently being field-tested.
"We are doing uncommon things because it's important for research to be relevant," said Dean. "Sometimes you do research because you're exploring." On other occasions, "you have technology in your hands that you want to apply, so you want to get Research tied into the outside world."
|
