Paper Reading #16, "A Practical Pressure Sensitive Computer Keyboard"

http://wkhciblog.blogspot.com/2011/03/paper-reading-15.html
http://chi2010-cskach.blogspot.com/2011/03/paper-reading-16-using-fnirs-brain.html

A Practical Pressure Sensitive Computer Keyboard

Paul H. Dietz, Benjamin Eidelson, Jonathan Westhues and Steven Bathich of Microsoft Corporation
Presentation venue not specified in paper, presumed to be UIST 2009

Summary
This paper is a report on research done in the area of an improved keyboard, in this case using pressure sensitivity to allow for increased features and versatility. Hopefully, this would all be achieved without paying a price in intuitiveness. For motivation, the authors lament that the current keyboard is still functionally the same as at 'the dawn of the computer'.

The authors follow by summarizing the prior work in this field, and explaining the technical details of the construction of current keyboards. These details are omitted, as I feel technical details are secondary and I want to focus on what this paper is doing, not what others have done before.

This keyboard is built as a 'modified flexible membrane design', with the goal of detecting increasing pressure only after the point where a conventional keyboard would have registered the key as pressed. I have also admitted the EE major's portion of the paper, because I don't understand and don't think I could do it justice. I will focus more on implications.

Suggested applications include gaming, where you could push a key harder to go faster, emotional IMs where font size scales with the force of key presses, and allowing minor increases in proficiency and accuracy in conventional typing.

Taken from the paper.

Finally, they discuss the potential obstacles to market viability and why they think this keyboard overcomes them. Specifically, cost increases are minor compared to conventional keyboards, no functionality is lost, and pressure sensing can be adapted by software allowing users to adapt slowly.


Discussion
While this approach has some merit, I think they're trying to force a niche a little bit. After all, the hammer and nail have been along for how long without being fundamentally replaced? The keyboard has remained unchanged because it does its job and does it well. The cutoff for sensing pressure being the point where a conventional keyboard would activate is the best idea in the paper.

This paper is interesting because of the potential applications, and significant because anything affecting keyboards on a wide scale will be very significant. That said, I have some serious concerns about the device as presented here.

I think they can be summed up fundamentally as concerns with moving from a digital control to an analog control, which would seem on the face of it to be backwards. I think that although it may be only marginally more expensive to produce, upkeep and replacement costs would increase as this keyboard would begin to lose precision well before the point where a current keyboard - where it either is pressed at any time or isn't - would finally die. Also, giving different results for pressing the keys harder would encourage users to be rough with the keyboards, further increasing these costs.

This lack of precision is pretty concerning to me as a gamer, I think. I don't emotional response from my controls, I want precision. I know exactly how fast I am going when I use any forward motion key, and I am concerned that it would make my movements awkward in games I already know to use such a device. Of course, that would probably just mean an adaptation period, so this isn't nearly as much of a concern as the above overhead.

One thought, which is mostly funny and not a criticism, since it could be easily corrected for, is that not all keys are conventionally pressed with the same force. Would J usually be larger than Q in the emotional IM proposed?

The biggest weakness of the paper is the lack of any testing, with users or otherwise, so far as I can tell. Accordingly, my idea for future research is to do extensive testing.

Finally, if the stenographers today don't use QWERTY, what do they use? Was the wording of this paper just misleading?