Hacker News new | past | comments | ask | show | jobs | submit login
Designs that Bell Almost Used for the Layout of Telephone Buttons (theatlantic.com)
112 points by Aqua_Geek on Sept 9, 2013 | hide | past | favorite | 69 comments



What I particularly like is the fact that, from a purely data-driven perspective, the obvious choice is I-C; the rotary-mimicing layout.

But (I conject) the lower error rate and preference is merely down to familiarity of users with the older rotary system.

So the lesson, for me anyway, is that listening to your users or to your data without applying your own reasoning can easily lead you to the wrong conclusion.

Which is not groundbreaking, but it's nice to have case studies.


*edit; it was lower keying time not lower error rate on I-C, but the point remains. I-C is the only layout with more than one special attribute.


Here is the link to the original BSTJ article.

http://www3.alcatel-lucent.com/bstj/vol39-1960/articles/bstj...


A friend is doing a PHD related to this subject.

Quite an interesting video of her talking about it if you are interested:

http://www.youtube.com/watch?v=kCSzjExvbTQ

Edit: Just realised Sarah's video is referenced at the end of the Article.


Numberphile in general is great. 5-10 minute chunks of interesting math. Not too long to be boring.


This title stretches the definition of the word "almost". They tested a bunch of varieties and settled on the best one. That doesn't mean the worst of them were "almost used".


Sometimes I think I have become so used to the extreme hyperbole of linkbait headlines that I forget it's even there. You're right.


Is there a browser extension that ensures I'll never see anything from The Atlantic, Slate, and Business Insider ever again? These sites are profoundly annoying and stressful.


You could edit your hosts file and map their domains to 127.0.0.1 or even filter them at your router, if it has that functionality.


You'll still see their articles listed on the front page. I think OP is looking for something that would remove it from the site altogether.


What about all Gawker sites? Probably worth adding those too.


I never consciously realized that the phone button layout was different than the calculator/number-pad layout. Weird! I'm curious whether I've misdialed numbers because of that.


I've lost my credit card due to that. ATMs have phone layout and i entered my pin using numpad layout three times T_T.


Came here to comment the same thing, I can almost blindly type these things out, but the change in the format has caught me out a few times.

My portable GPS unit by default has the ABC layout keyboard, that is just as annoying to use when you are used to QWERTY!


This causes frequent but barely-traceable errors in math classrooms, in my experience.


It's really odd to me that the traditional keyboard 10-key entry pad hasn't switched to the phone number layout. If the phone layout is more efficient and less error prone, why didn't keyboards start with the phone layout and not the calculator layout?


For 10 key entry in accounting you want the most often used numbers closest to you. Benford's Law explains why 1 is the most common digit, followed by 2, 3, etc. Thus for hand entry, small digits should be closest to the hand, at the bottom, and larger digits at top.

For dialing a phone you are not constantly dialing numbers as a accountant or data entry person will be doing. For phoning, the number 1 is not the most common one, it's was actually more rare in old numbers which reserved 1 and 0 in some positions for area codes, needed only for long distance. The user testing showed that top to bottom order worked better for users than bottom to top. Undoubtedly since it matched reading order of the english language, going top to bottom as well.

The most efficient and least error prone layout depends on the task. The task of accounting, and the digit frequency, is very different from the task of telephoning.


The digits 1, 2, and 3 appear more often in data than the digits 7, 8, and 9, so you can do data entry faster with the existing layout.


http://www.vcalc.net/Keyboard.htm has lots of theories. I had a vague memory that it was due to some sort of patent issue but given a complete lack of verification on that I distrust that memory.


Might only be less error prone for entering phone numbers.


I was thinking more along the lines of early computer only really being used by people who had a stronger muscle memory for calculator usage than telephone usage.


It's interesting to think about how the use of one of these alternate layouts would influence today's phones. I wonder what the iOS/Android phone apps would look like with a circular button layout.


There was at least Nokia 3650 with something like I-C; I must say I liked its layout but Nokia never continued it.


This makes me wonder how many HackerNews readers have actually used a rotary phone.


It's like remembering the screeeeee-krrsssssh-bdang!bdang!bdang!-krrrssssh-click of a dialup modem. It fairly neatly bifurcates generations of nerds.

Rotary phones were quite elegant. As the dial returns to 0, it sends pulses down the line. The pulses actuate series of stepping switches to form the final phone circuit.


> Rotary phones were quite elegant. As the dial returns to 0, it sends pulses down the line. The pulses actuate series of stepping switches to form the final phone circuit.

Back when I was poor and didn't have a place to live, I would go into a hotel lobby where they had a phone for guests to receive, but not make, calls -- the phone didn't have a dial. I would pick up the handset, wait for a dial tone, then press one of the buttons beneath the handset, ten time really fast. Then I would say, "Hello, operator? I'm handicapped, can you dial a number for me?"


Australia's phone system largely skipped past having operators, so the various tricks for getting one to dial for you never worked here.


Ah, but with a little practice (& patience!) you can dial complete numbers using this system. I used to call the 113 number (clock/time information) here in Argentina in this way just for fun.


Just as well. Human operators only prolonged unfair and very negative sexual stereotypes.


It's ok, we're Australians. We have plenty of other negative sexual stereotypes to pick up any slack.


It makes me wonder how many have re-mapped their phone buttons to whatever the Dvorak equivalent of the numeric keypad would be. With blank keys, of course.


Blank keys on a touch screen phone! You invented the ultimate affectation.


I remember, as a kid, wondering whether the clicking of the rotary dial could be simulated by clicking the hook button rapidly. It worked! I could dial the phone by simply tapping the hook button.

At one point I tried connecting a simple speaker (from a little transistor radio I had taken apart) directly to the phone line. I got a dial tone. By connecting one wire to the speaker terminal, and tapping the other wire on the other terminal, I could dial a phone number. Those old phones were really very simple things.

No idea if the phone network still supports pulse dialing. I'd guess it does, but haven't tried it.


I've used a rotary phone, on a party line. (https://en.wikipedia.org/wiki/Party_line_(telephony))


I stayed in a hotel in Hiroshima in 2011, and my room had a rotary phone (the room generally looked like nothing had changed changed since about 1971)! :]

[The actual hotel: "アレーホテル広島並木通". It was actually a very nice hotel—clean and neat, friendly service, free (if simple) breakfast, crazy-low prices, and an incredible location—just a bit out-of-date. I highly recommend it for anyone wanting a budget hotel in Hiroshima. They also have slightly more expensive rooms that look a bit more modern (maybe even rocking a touch-tone! :).]


Haha I had the hardest time guessing what 外来語 「アレー」 came from. I thought it was the French "allez" at first, but it turns out it's "alley"? That's usually written 「アレイ」.

And speaking of outdated technology in Japan: http://www.nytimes.com/2013/02/14/world/asia/in-japan-the-fa...


I have. I don't recall where I first saw it, but I do recall thinking to myself, "What's wrong with this thing? It must be broken -- it is making clicking sounds instead of a proper beep."

Shortly after ward, the purpose of those pulse/tone switches on more modern touch-tone telephones dawned on me: they existed for backwards compatibility with especially old phone networks which didn't support touch-tone devices, only rotary-dial ones. That such networks actually existed and were still in operation didn't seem such like such a crazy idea, considering the museum piece I was holding in my hands.


Probably more than those who used the letter prefixes. I remember our 527-xxxx was JA7-xxxx. I don't remember ever telling anyone the "JA7" version of our home number, but it was the way we remembered our doctor's number. I feel old.


I'm 31 and my family home had nothing but a rotary phone until ~1995. We didn't get a TV with a remote until around the same time, and it only got 3 channels. My parents still only have 3 channels living there.

I grew up in Australia :)


I wonder how long the telephone number will be around for? Another 10 years? 100?

Do you think 1000 years from now we'll still be using them?


Can you rattle off your facebook account number? Chances are, few people are more than semi-consciously aware that such a thing exists.

Chances are the phone number will remain, but it won't be important ... like Facebook, and like modern phones, you deliberately connect with someone. True with modern phones you do consciously give them a number whereas with Facebook you don't (even though it's the underlying account id #'s used by Facebook's infrastructure), but there are other options such as bumping someone's phone with yours to initiate transfer of some piece of data or other.


I would like to think it'll soon be slowly replaced by something more sophisticated.

Personally, I'd love to have control over who can contact me and by what means - basically, grant access to specific pieces of my contact information to specific third parties where I can revoke access at any time and they get any updates I make instantly.


I'm struggling to find a difference between IV-A (the winning design) and VI-A.

Is there a subtle difference? They look the same to me.

If they are the same, why would they have two identical designs?


Saw this too. I suspect each row refers to a specific head-to-head test of 3 designs. If so, you should only compare statements within each row, and not vertically.


Keying Time. Look at the asterisk next to IV-A (significantly shorter keying time)


I remember having a touch-tone phone with the buttons in the positions of the finger holes for a rotary-dial phone. I think it was supposed to look retro?


I hope that the cross pattern came with gospel sounds instead of the standard beeps.


They should have put the zero before the one.


Yes!! It's remarkable that only one or two of the designs had the numbers in order.

I've always wondered about this in regard to keyboards as well: why did anyone ever think it was a good idea to put one of the numbers out of sequence?


Zero was (is) special. It dials the operator. This was much more common back then than it is today, so it is fitting that it got a special position. I can't speak for keyboards, but I do recall once reading a Knuth paper on METAFONT where he noted the difference between the order that mathematicians write numbers (0123456789) and the order that typographers write numbers (1234567890).


Are you aware that for a long time many typewriters didn't even have a one key? You were supposed to type a lowercase ell for one. And if you really needed an exclamation point, you overstruck a period with a single-quote.

Obviously that could never had worked on a computer keyboard, and I think the trend to supply an actual one key was in motion before keyboard-enabled computers became common. My point is that the folks who invented and (initially) used typewriters were not thinking mathematically. They were 100% concerned with getting ink on paper quickly and legibly.


In case of rotary phones it was much easier to put zero at the end. The old phones used pulses to dial numbers - one pulse meant one, two pulses - two, and so on. Ten pulses meant zero. If they wanted to put zero at the beginning, the dialing mechanism would need to be much more complex.


In Sweden, one pulse was (is) used for zero, two pulses for one and so forth. Therefore, old rotary phones havet zero next to one [1]. However, when buttons were introduced de ended up with the same layout as the US.

1:http://upload.wikimedia.org/wikipedia/commons/f/f9/Dialog.jp...


At least on the numpad it makes sense. One of the most used buttons in accouting (the 0) is assigned to your thumb and it can rest on the oversized button.

As for the number row, this seems to sum things up nicely: http://ux.stackexchange.com/questions/39544/keyboard-number-...

Basically the 0 was added near the O because pre-zero being included typists used the capital O as zero. Thus zero was placed near O as a matter of usage.


Because non-programmers start counting at 1. 0 is only used when you get to 10.


0 is 10 pulses on a rotary dial. 0 is not really a number technically is it, more of a digit?


Zero is not technically a number? It has been a long time since my grade school days, but I seem to remember zero being an element of Reals, Rationals, and Integers. Granted, zero is not an element of what we used to call "Counting Numbers," but really, it should have been.


None of these include the ABCD keys (FLASH OVERRIDE, FLASH, IMMEDIATE, & PRIORITY) used by AUTOVON which started in 1963.

It's interesting that all of these layouts are presented on a round background.


Book Recommendation: "Design of Everyday things". A great book,and particularly covers quite a bit about telephones.

Surprised no one has brought this up yet. :-)


I like III-B.

It seems to have the least amount of finger-roam of all the layouts.

Yet it offers more keying discernability than VI-A owing to the tiered layout.


I wonder if they took the visual mnemonic shortcut we use to remember the numbers into account. That would explain why they settled on a uniform almost square grid.


And V-A would have supported "home row"-style touch-typing for all 10 numbers, plus room for the extra two buttons that commonly get added.


Based on this sheet, it looks like they decided on the layout we use but wanted testing to back it up.


Hm, IV-A and VI-A is the same?


From the comments:

2A and 6B are the same too. Maybe the repeats got the best results in the first five sets of trials so they brought them back and pitted them against each other as well as the traditional rotary phone number arrangement in round six. 6A eventually won, obviously.


These were tests. They tested each group against each other, so some of the patterns were tested more than once.


Maybe. It looks like the buttons are a bit smaller with slightly more space between them.


Aren't VI-A and IV-A the same layout?


Ya. And aren't II-A and VI-B the same? It bothers me that they didn't mention anything about this.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: