QWERTY Must Die, by Bryce | Hacker News Reader

QWERTY Must Die, by Bryce

bryce.vc

Using "QWERTY" as the noun here instead of "typewriter-like keyboards" is confusing everyone. He's arguing against physical keyboards, not a specific keyboard layout.

(Typed in Dvorak)

I find that overly optimistic that qwerty will go away by the time his 18-month-old is an adult. QWERTY has been with us for over 100 years, and I see little need to throw it out.

It is a good, functional, accessible, and consistent tool for putting words to screen. I can't understand why so many techies are so eager to throw everything out the window.

QWERTY has no rhyme or reason and must be laboriously memorized (and I say this having reached ~80 WPM years ago). It's even been claimed that it's deliberately de-optimized for our natural languages to accommodate legacy hardware (pre-Selectric typewriters). If gestures are the future, text entry should take advantage of the flexibility of virtual keys and offer better Fitts' Law targets for more frequent letters or digraphs. I like Swype well enough but I can imagine something derived from pie menus being a big step up.

Yes, it is laborious and must be memorized -- just like our multiplication tables back in 2nd grade. Learning is not always fun, nor should it be.

Yes, but the difference is that multiplication is a fact about logic, not a contrived interface to obsolete hardware. Multiplication tables do have rhyme and reason to them. We don't memorize our 21-times tables because as soon as we pass 9, we switch to a methodical system with a simple, general system of rules.

It is pretty hard to swallow that something developed a hundred and forty years ago for interacting with box of springs and levers is a close-to-optimal text entry system.

Any keyboard will be a "contrived interface". The logical thing would be alphabetical order for ease of learning, but that would be terrible for speed. Alternatively, you could go with something like Dvorak, optimized for speed, but at the end of the day Dvorak feels just as contrived, has the same awkward learning curve, and the evidence that it's significantly faster than qwerty is virtually nonexistent except for a couple of old studies whose results have not been substantiated by newer studies.

The point is not that "contrived = bad". The point is that a contrived interface can be optimized, while multiplication tables are a fixed part of understanding multiplication in a positional system.

On the other hand, you could argue that we could optimize our entire number system for learning by switching to a lower base. And in this sense I do agree that optimizing for ease of learning would be a mistake. If we taught base-3, learning multiplication tables would take one tenth of the time, but actually multiplying numbers would a lot longer. Analogously, I would gladly spend 10 times as long to learn a text entry method if it made me 20% faster in the long run.

So I guess what I'm saying is that QWERTY should be replaced not because it's hard to learn, but because it can't possibly be optimally efficient.

Dvorak is not faster, but it is significantly more comfortable.

At least, that's what a lot of people say. I'd love to see some decent research. I guess it's hard to double blind for things like keyboard layouts.

You could get a bunch of qwerty users/untrained typists and then train them to use randomly assigned keyboard layouts. Once training is complete, you have them type some long piece of text, and then fill out a questionnaire rating various qualitative points like tiredness.

Finally, you do a statistical analysis to determine if layouts with more of Dvorak's purportedly beneficial features are correlated with higher comfort ratings. Of course, those features must be rigorously defined beforehand, and you must properly blind yourself when doing the analysis.

There were a lot of things developed more than 140 years ago that we still use. For example, let's take the rocket engine, developed by the Chinese in the 13th century. I believe a very similar engine was used to put man in space.

What is old is not necessarily antiquated.

My point is not that what is old is antiquated. My point is that there were a tremendous number of constraints on the 140-year-old design of the typewriter-style keyboard interface that simply do not exist today. Even if the typewriter's interface were optimal under those constraints it would be very unlikely that it would be optimal once those constraints were lifted.

That makes it a perfect fit for language. The alphabet, spelling rules, and grammar aren't much different, unless, of course, you immediately recognize a lowercase 'a' because an 'A' looks like an upside down oxhead, and 'ox' starts with an 'a' in early Greek (or was that Sumerian?)

I do not think any extra mental load that learning qwerty adds is significant.

Do piano players ever discuss rearranging the keys + pedals on a piano?

(Not troll bait. Honest question.)

The piano keyboard is actually pretty well-designed with a very consistent design. Every key (both black and white keys) is a half-step from each other. Learning scales sucks initially, but once mastered, the dance is easy and fun.

I've often wondered this too, a "Dvorak for pianos".

Googling it now turns up: http://en.wikipedia.org/wiki/Wicki-Hayden_note_layout

Yes , two categories :

http://en.wikipedia.org/wiki/Isomorphic_keyboard These tackle the problem of how the normal piano layout requires different motor-memories for the same interval-patterns at different pitches . Janko and Wicki-Hayden are examples , as is layout of the chromatic button accordion .

http://en.wikipedia.org/wiki/Generalized_keyboard These are designed to work for various other tunings than standard 12-tone tuning ; i think most of them map some pitch-space to a hexagonal grid .

The strongest argument he makes is that there's an impedance to text entry vs. "tactile computing." Which seems pretty intuitive to me. With a keyboard a user can generate tons more information (as opposed to raw data) than with almost any other input hardware.

But you have to think about what information you're inputting. A GUI provides an ongoing dialog that guides you through the interaction with the computer. But if there's data that's unavailable a priori to the interface developer, a keyboard is still the best way to enter it.

> For the longest time I thought the answer would be speech to text, but the more I play with Siri and work around its shortcomings the less and less of a believer I become.

Really? You think Siri's (current) technical shortcomings are the major hurdle for speech-to-text adoption?

Not the fact that no one wants to type out loud in their cubicle or coffee shop?

There are some specialist niches for speech to text. Medical notes are one example.

I'm gently surprised that speech to text isn't better. There's so much weaponisation of sound analysis that I thought the research would be there to help programmers.

Also, I don't know where you drink coffee but there's a total arsehole in one local coffee shop who's happy to bellow into his mobile phone.

And let's all switch to Esperanto while we're at it.

dvorak

Reply on news.ycombinator.com