Triangular buttons key to touchscreen typing success | Hacker News Reader

Triangular buttons key to touchscreen typing success

reghardware.co.uk

18 pointskqr222 comments

So he's trying to patent the use of triangles for keys? I wonder who nailed that lucrative square key patent. He must be worth billions by now.

Patenting qualms aside, I think this is a brilliant idea.

I wonder why this hasn't been implemented. I could definitely see using this format for detection, but then visually using square buttons. Best of both worlds?

Agree.

It's not pretty and round.

It doesn't make any sense to me. On a touch screen, you would have the touch point and you can see which key center it's closest to.

Here's my suggestion for improving touch screen keyboards:

http://news.ycombinator.com/item?id=519052

> I could definitely see using this format for detection

Why have any gaps between the keys at all?

On a physical keyboard, you want gaps for at least a couple of reasons: so that your fingers can feel the edges, and so that you don't accidently click on two keys at once. Triangular keys might make some sense on a physical keyboard then, since presumably you would be able to press inside the combined area of the key's triangle and the two adjacent triangular gaps, so less chance of pressing two keys at once.

On a soft keyboard, you can't feel the edges of the keys, so the first point doesn't apply anymore. The second worry, pressing two keys at once, also doesn't apply, since the software calculates the centre point of the finger press, and activates the key which that point falls over, rather than all the keys under your finger.

With a soft key pad then, the only real way to make keys easier to press is to make them bigger. The iPhone for example, will increase the effective size of the keys that you are likely to press next, predicted based on language settings.

So, it probably hasn't been implemented for a soft key pad because it would make it harder to press the keys. You would press roughly in the right place, and get no response.

As for physical key pads, I notice that the Blackberry has keys with raised diagonals. That probably gets you most of the advantages of triangular keys (easier to feel them apart) without the need to actually make them triangular.

http://3466.voxcdn.com/wp-content/uploads/2007/09/8820-keybo...

Agreed. This is why I found it bizarre that the BlackBerry Storm's on-screen keboard mimics the 2-letters-per-button design of some of its hardware keyboards.

Pic: http://www.infoworld.com/infoworld/img/48TC-storm-portrait.g...

The buttons are all virtual, and the only thing that matters is the (x,y) coordinate of the screen press. so why encourage users to press the screen somewhere other than the center of the desired letter?

I'm imagining that the real HCI difference here is that which key your finger is determined to be centered on changes. Whereas before, touching the "gap" between a key above and below your finger would have an unpredictable 50/50 chance of activating either key, now the "gradient of precedence" of the keys fade out as they near the triangle point, so, assuming the triangles were pointed downward, you would always activate the key below. Making things predictable makes users happy. You could do this without the triangular display, of course; it's simply an interesting affordance to hint to users that they can use the keys this way.

Wouldn't that be the same as shifting the active area of the keys a few pixels upwards?

Dead space would just be assigned to the nearest possible key anyway so this isn't much gain.

Now, if you had tessellated triangles, that might work. Right now, the keyboard works by row and column. I suspect it's more common for the iPhone to trigger "f" as "d" or "g" rather than "r" or "v": in other words, detecting side to side is harder than up and down. That suggests to me that if you tessellated triangular keys together, how far up and down you are would help determine whether you were going for an upward-pointing key or a downward-pointing key. This isn't very clearly stated but I hope it makes some sense.

Tessellated hexagons might be interesting - they have the advantage of being closer to a circular target.

Apparently I'm not the first one to think of that - a quick check threw up http://www.freepatentsonline.com/5805157.html

It's easy to make the keys look far apart in a landscape mode. Go back to portrait and talk to me then.

Something about this says "klingon" to me, but I can't find a good screenshot of the UI I'm thinking of.

You can actually buy a ps/2 compatible Klingon keyboard although the keys are not triangular:

http://www.cherrykeyboardsrus.co.uk/Klingon+Language-Details...

This sort of makes sense to me for small tactile keyboards, but with a touch screen, why does it even matter? You could just as easily have no gap between keys, but calculate a "confidence" level of what key was pressed based on how close to the center of the key the press occurred. It seems to me that this layout would just turned incorrect key presses into ignored key presses if you hit in the dead space around the triangles. I guess you can consider that better since you don't have to backspace, but a decent smart-correction implementation is better still.

Not only that but you should be able to add predictive text logic to the decision process: pick the next letter both based on the location of the key press and the next letter that we think you would type anyway. If you last typed a Q you probably next want a U not a J.

Edit: DougBXT claims below that the iPhone already does this.

I believe the iPhone does do this. Because of this, what is the point of having the triangular keys? I suppose changing the onscreen appearance of the keys may cause people to touch more toward the center of the key, but that would only yield better results at the expense of typing speed.

I googled for his name and, as far as I can tell, he doesn't have a web presence, he's not an academic or has published an academic paper and he doesn't have any quantitative studies up backing up what he's done.

Not saying it won't work, just saying that he seems like an outsider and so I'm going to be sceptical unless he demonstrates something a bit more substantial than a puffed up media piece.

I'm skeptical of claims without any research to back it up. But some thoughts (DISCLAIMER this is BS)

An effect of having triangular keys instead of square keys on a soft keyboard is that each key has a clearly delineated boundary rather than sharing boundaries. I wonder if that can have an effect on how well people can associate a given shape with a letter when looking at it.

I also might guess with the extra space it takes longer to visually scan across the keyboard if you are looking for a key. So if there was a benefit to accuracy maybe it would be from slowing down the user enough for them to really process what they are looking at.

Baker told Register Hardware today that each triangular key has significantly more dead space around it than you’d find on a standard Qwerty layout. Consequently, users are more likely to press the correct key each time they tap.

Or less likely to press any key at all?

This could be made into an iPhone app for email/sms.

It wouldn't replace the native keyboard on the iPhone, but would allow users to compose their message using this triangular keyboard and forward the text to the appropriate application.

I'd like to see quantitative studies showing how much of an improvement this is! That should be very doable with the App Dev Kit.

Reply on news.ycombinator.com