Haptic interfaces
in [Embedded]
|
From: Tim Wescott on 1 Mar 2010 13:19 Boudewijn Dijkstra wrote: > Op Mon, 01 Mar 2010 04:44:29 +0100 schreef D Yuniskis > <not.going.to.be(a)seen.com>: >> What makes a haptic interface "good"? > > Dunno. > >> bad? > > When it is painful. ;) Actually, if it were on the primary controls of a plane, something that gets painful when there's a good chance of damaging your life-preserving airframe 10000 feet above the ground may not be a bad thing. Particularly if there's enough pain to make you think, but not enough to force you to fly into Mt. St. Helens on a particularly bad day. (This is from a criticism of flight control systems, by the way -- at least at the time it was made several years ago, Boeing automatic flight controls were criticized for letting you bend the airplane, while Airbus flight controls were criticized for not letting you keep the plane from flying into terrain. I don't know if things have changed since then.) -- Tim Wescott Control system and signal processing consulting www.wescottdesign.com
From: D Yuniskis on 1 Mar 2010 14:50 Hi Boudewijn, Boudewijn Dijkstra wrote: > Op Mon, 01 Mar 2010 04:44:29 +0100 schreef D Yuniskis > <not.going.to.be(a)seen.com>: >> What makes a haptic interface "good"? > > Dunno. > >> bad? > > When it is painful. ;) Wow, I never considered that possibility. :-/ >> *Exceptional*? > > Perhaps a lot of positive scientific papers about it. <frown> >> What would you nominate as the "best" haptic interface? > > For doing what? Pointing something out on a screen, inputting text > and/or context-sensitive instructions, digging a ditch, positioning > concrete slabs, removing a kidney and piloting an aircraft would usually > benefit from different interface properties (haptic or not). I disagree. I am sure there are attributes of a haptic interface (or any other interface, for that matter) that make them better or worse than other implementations in their class. >> What (electronic) device would you nominate as having the >> best haptic interface? > > What is the difference from the previous question? Why does an interface have top be part of an electronic device? E.g., I would nominate a traditional "squashed sphere" doorknob as the best haptic interface. It's shape ("feel") is reasonably comfortable (not painful nor intimidating -- hence the qualification of the "squashed sphere" variety and not some of the more exotic artsy-fartsy door handles), it doesn't "prefer" a particular size hand, it is intuitive in operation, can be operated without benefit of any of the other senses, etc. Contrast this with something like a "child proof" pill bottle... the size of the cap varies and, as such, causes it to favor a particular hand size/strength (arthritic senior citizens vs men with "manly" hands vs teenagers and, of course, young children); the childproofing usually makes the actions required to "unlock" it counterintuitive (on purpose), etc. Note the criteria used in these explanations identify the sorts of things that I perceive as "important" (to *me* -- the purpose of the question was to identify criteria that others might consider important). I, for example, consider most of Apple's (electronic) offerings to have crappy interfaces. You *need* your eyes to use them (even on things like setting the volume on an iPod), they require more attention than should be necessary (again, the iPod example comes to mind -- changing volume should be the sort of thing you can do without thinking about what you are doing -- instead of requiring you to track your finger in a particular circular orbit on the face of the device -- note that things like the Shuttle were much easier to operate in this regard); they aren't particularly intuitive (contrast with the doorknob which even a toddler can "operate"), etc. Refering to your "For doing what?" question, could you consider this sort of iPod interface BETTER in *any* application than some other haptic interface -- ignoring trivial cases? (this is a genuine question, not a statement of my beliefs) >> And, of course, "why", in each case.
From: D Yuniskis on 1 Mar 2010 14:55 Hi Tim, Tim Wescott wrote: > Boudewijn Dijkstra wrote: >> Op Mon, 01 Mar 2010 04:44:29 +0100 schreef D Yuniskis >> <not.going.to.be(a)seen.com>: >>> What makes a haptic interface "good"? >> >> Dunno. >> >>> bad? >> >> When it is painful. ;) > > Actually, if it were on the primary controls of a plane, something that > gets painful when there's a good chance of damaging your life-preserving > airframe 10000 feet above the ground may not be a bad thing. > Particularly if there's enough pain to make you think, but not enough to > force you to fly into Mt. St. Helens on a particularly bad day. Wow! That's a great idea! I.e., not just "feedback" but "particularly unpleasant feedback" that really works to dissuade you from doing something that you shouldn't. I've designed "big knobs" with force feedback (to simulate the "mechanisms" you are influencing with your "adjustments") but those just gave you subtle reinforcement that you are "doing what you expect to be doing" (e.g., if the knob is supposed to cause something to be elevated, then it is harder to turn in the "up" direction than the "down" direction). But, that raises the issue of "what happens if you (the device) screw up" and your "penalty pain" makes it hard for a user to "do what is right"? (think of this in the example you cite below; I've heard avionic controls are far from "perfect" :> ) > (This is from a criticism of flight control systems, by the way -- at > least at the time it was made several years ago, Boeing automatic flight > controls were criticized for letting you bend the airplane, while Airbus > flight controls were criticized for not letting you keep the plane from > flying into terrain. I don't know if things have changed since then.)
From: Tim Wescott on 1 Mar 2010 14:47 D Yuniskis wrote: > Hi Tim, > > Tim Wescott wrote: >> Boudewijn Dijkstra wrote: >>> Op Mon, 01 Mar 2010 04:44:29 +0100 schreef D Yuniskis >>> <not.going.to.be(a)seen.com>: >>>> What makes a haptic interface "good"? >>> >>> Dunno. >>> >>>> bad? >>> >>> When it is painful. ;) >> >> Actually, if it were on the primary controls of a plane, something >> that gets painful when there's a good chance of damaging your >> life-preserving airframe 10000 feet above the ground may not be a bad >> thing. Particularly if there's enough pain to make you think, but not >> enough to force you to fly into Mt. St. Helens on a particularly bad day. > > Wow! That's a great idea! I.e., not just "feedback" but > "particularly unpleasant feedback" that really works to > dissuade you from doing something that you shouldn't. > > I've designed "big knobs" with force feedback (to simulate > the "mechanisms" you are influencing with your "adjustments") > but those just gave you subtle reinforcement that you are > "doing what you expect to be doing" (e.g., if the knob > is supposed to cause something to be elevated, then it > is harder to turn in the "up" direction than the "down" > direction). > > But, that raises the issue of "what happens if you (the > device) screw up" and your "penalty pain" makes it > hard for a user to "do what is right"? (think of this > in the example you cite below; I've heard avionic > controls are far from "perfect" :> ) That's why I was calling out a level of discomfort that could be overcome with enough adrenalin. To pull an example from real life, if you're out jogging and you pull a muscle you should slow down and walk -- but if you're getting chased by a dozen toughs and you pull a muscle, you should run like hell anyway! I wouldn't put the discomfort level up to 'painful' unless it really was a life-threatening problem. But fly-by-wire systems already have 'stick shakers' to warn the pilots that they're pushing the envelope of safety. -- Tim Wescott Control system and signal processing consulting www.wescottdesign.com
From: Joel Koltner on 1 Mar 2010 15:08
"D Yuniskis" <not.going.to.be(a)seen.com> wrote in message news:hmh540$lr1$1(a)speranza.aioe.org... > E.g., I would nominate a traditional "squashed sphere" doorknob > as the best haptic interface. ....but this is often cited as as very poor interface for those with limited dexterity -- hence the trend to have handle-shaped door openers. (Here in the U.S., on most commercial buildings the ADA makes it illegal to even have round door knobs.) > Contrast this with something like a "child proof" pill bottle... Indeed. Some pharmacies (e.g., Target) have also taken on the problem that reading a label on a cylindrically-shaped bottle is rather less than idea, and now use kinda tapered oblongish bottles with flat surfaces for the labels. > I, for example, consider most of Apple's (electronic) offerings > to have crappy interfaces. I think they're quite intuitive, although of course what one finds intuitive is often largely a function of what they already know. :-) And I agree with you that mechanical jog dials are better than touch-based ones. It's true that Apple's interfaces are often not designed to be used single-handedly/without looking at them/etc. -- this is a trend in all of electronics and software to make things "novice friendly" often at the expense of them no longer being "expert friendly." ---Joel |