The Birth of MultiPoint

It’s incredibly exciting for me to see the creation of this blog and what it represents – increased momentum behind MultiPoint for PCs in schools!

MultiPoint was conceived during the summer of 2005. That summer, a then-PhD-student from UC Berkeley, Joyojeet Pal, did a research internship with us at Microsoft Research India (MSR India) to investigate how computers were being used in 30 rural primary schools across India, most of which were sponsored to some degree by the Azim Premji Foundation. In his final presentation, he showed pictures of PC after PC in schools... and what was interesting was that, unlike in wealthy schools in developed countries, they were each surrounded by several kids. In some cases, two or three students sat at one PC, in others, as many as ten. In most of these cases, a dominant bully hogged the mouse and keyboard. Joyojeet, moved by the injustice of these arrangements, recommended that teachers move students around periodically, so that different children would have a chance to interact with the PC.

When I saw the photographs, however, my first reaction was to fix the problem with technology (technological fixes, I’ve since discovered, are not always a good idea, but it happened to work in this instance!): If you have five children at a PC, why not provide five mice? Unfortunately, at the time, I was too busy with other work to follow up.

That August, we were fortunate to be joined by Udai Singh Pawar, a fresh graduate of IIT Kanpur. The Indian Institutes of Technology (IIT) are renowned worldwide for their science and engineering graduates. When Udai came on board, I gave him time to think about what he might want to work on, and suggested a few possible research topics including "multi-mouse." Udai was a physics major and had worked with novel keyboards and robotics-for-children as an undergrad. Building things and working in education were in his blood. And, after a couple of weeks, he told me, "I’d like to see what we can do with multiple mice."

Udai worked fast. Within a couple of weeks he had a demo of several cursors dancing on the screen, each controlled by a mouse. There were glitches in the early prototypes – the cursors blinked in and out or they adopted non-cursor shapes – but I was nevertheless amazed and delighted. He then started planning a series of studies that would allow us to observe how real students would respond to multi-mouse, and also wrote a couple of simple educational applications he could use to test. One of the applications was a quiz game in which players hear a letter of the Kannada alphabet (local language in Bangalore, where our lab is based), and then must click on the correct letter among several.

We started with an initial set of informal studies, just to get a sense for whether the concept would work. Would kids get it? Could they distinguish their cursors from others? Would they like it? We tried various groups of children first interacting with the game with one mouse for a while; then, after 10-20 minutes, we’d add mice to match every child. It didn’t take us long to discover that the children loved it! In fact, a few told us they didn’t understand why all PCs didn’t come with lots of mice. We tried this with several groups of children, and when we were convinced that this was worth testing further, we tried a more rigorous evaluation.

We wanted to show that MultiPoint could be as effective as a single-PC-per-person configuration for at least some learning tasks. We decided to test something that had some basic educational value in India and was also easy to measure: English-language vocabulary. We set up a game where the goal was to match pictures of animals with their English names – e.g., cow, elephant, snake, ferret. There were two versions of the game – competitive and collaborative – with both versions “collapsing” to the same one-mouse version of the game with there was only one mouse. We put students in groups of five, with various ratios of boys and girls. They were then administered a pre-test which tested what words they already knew, allowed to play the game for 15 minutes, and then given a post test. All together, 238 students participated in the evaluation.

The results showed that on the whole, students could learn as much with MultiPoint as with a single PC to themselves. This was great news, because it meant that at a fraction of a cost of a single PC, students could learn as much in certain types of educational activities: With five students at one PC, the per-student cost of computing is approximately 1/5 that of a one-PC-per-child configuration, which many schools cannot afford. In addition, there were gender effects that researchers of technology and education have long known about. Boys always want to compete, and will compete blindly without learning much, unless the game forces collaboration. Girls always seemed to help each other regardless of the set up. So, a collaborative game is preferred to tilt the odds in favor of real learning for everyone. You can read more about this and the initial research for MultiPoint at   

The rest is history… When we showed these results to Sherri Bealkowski, then a general manager in Microsoft’s Unlimited Potential Group (UPG), she immediately saw the value for schools, and worked with Jed Rose and Microsoft’s Education Product Group to develop a free MultiPoint software development kit (SDK). Meanwhile, Neema Moraveji from MSR Asia heard about MultiPoint and proposed to apply multiple mice to a whole-classroom situation, which he did with much success in China (he's now a PhD student at Stanford). Bill Gates now speaks about MultiPoint as a kind of “creative capitalism,” where corporations spend a portion of their resources to design products and businesses for developing markets. And, most recently, Nasha Fitter joined UPG as a senior product manager to spread MultiPoint wherever it might be useful. This blog was her idea, and I couldn’t be happier that she’s joined the MultiPoint virtual family. Maybe you will, too!


Comments (2)

  1. nickboss says:

    I try to create a MultiPoint control using the template provided with MultiPoint SDK, however I don’t seem to make it.

    Is there any tutorials or examples of a MultiPoint control, except MultiPointButton, so that I can use?

Skip to main content