Show HN: iPhone 2005 weird "Blob Keyboard" simulator

129 points by juliendorra 1 year ago | 54 comments

Hi HN,

I teach tech design history, and one of the key stories I cover is the development of the original iPhone keyboard by Ken Kocienda. Reading about it in his book "Creative Selection" is great, but I wanted my students (and now you!) to actually feel this step in the process.

So, I built a web simulator of the "Blob Keyboard", Kocienda's very first attempt at a touchscreen keyboard that actually works, from September 2005:

Try the Blob Keyboard: https://juliendorra.github.io/blob-keyboard-simulator/blob-k...

- Tap for the middle letter

- Swipe left or right for the side letters

More on the github repo: https://github.com/juliendorra/blob-keyboard-simulator

The Blob Keyboard prototype emerged during a UX crisis for iPhone team (their software keyboard just didn't work at all, fingers being too big, and the Newton failure loomed over them), highlighting how innovation is rarely a straight path. It was developed on a tethered touchscreen display codenamed "Wallaby".

To make this simulator as authentic as possible, I referenced images from Kocienda's book and even got direct feedback and guidance from Ken Kocienda himself on Bluesky.

What to expect (or… what not to expect):

This is a reconstruction of a very early prototype with limitations reflecting that specific moment. The goal was to test first if typing with accuracy was even possible, as all the rest was moot if it failed!

It's NOT QWERTY: They were still hoping to get us out of QWERTY, but then familiarity won.

No Backspace: You can't delete.

No Cursor Movement: The text field is just a simple display.

No Caps or Numbers: Only lowercase letters.

No Smooth Animations: Keys just "pop" instantly when pressed. Kocienda noted that your eye fills in the gaps, giving a sense of movement.

Best Experience:

While it works with a mouse/trackpad on desktop, it's designed for touchscreens to better replicate the original Wallaby hardware interaction. Use it on your phone!

This project aims to provide a tangible glimpse into a turning point moment in iPhone development and the iterative nature of design. It's like stepping back in time and trying out that early demo on Kocienda's desk.

I would love to hear your reactions and thoughts on experiencing this piece of UI history! What other significant prototype do you wish you could experience?

bryanlarsen 1 year ago |

Does the book discuss capacitive vs resistive touchscreens?

At the time you basically had two choices for touchscreens: resistive or capacitive. Resistive was "obviously" the way to go because it was far more accurate. Choosing capacitive was inspired -- when used with stubby fingers the accuracy problem was moot, and it allowed multi-touch.

Just before the iPhone came out I was fairly confident I knew what the future was. It was now possible to create a phone with the horsepower to run a real web browsers. 800x600 pixel screens were available which would display normal web pages nicely, and a resistive touch screen with a stylus would make them useful.

Then the iPhone came out. 320x480 screen meant normal web pages wouldn't display properly, inaccurate touchscreen meant tap targets needed to be increased massively. Why would anybody buy an iPhone which didn't allow you to install apps, and the web was unusable because it required rewriting every page since existing pages were unusable. Instead you could buy a phone which allowed you to install apps and which allowed you to usably access the web. Obviously the iPhone would be a failure. :)

juliendorra 1 year ago | |

Yes, the book explains how everything started from the capacitive touchscreen. The initial idea (2004-2005) was to build a Mac tablet computer based on touch screens. Bas Ording designed all the interactions we know, rubber band and inertial scrolling, the home screen… for a mac tablet!

So really the capacitive screen drove the interactions. Input first, just like the mouse on Macintosh or the stylus on Newton, everything then flows from there.

On the web browser, I disagree with you (sorry!), the killer app of the iPhone was that Safari was the same Safari, with the same capacity and rendering, than on desktop.

It was completely new. Yes, you had to double tap on complex, non responsive websites, but every single (non-flash) site would render the same.

My 640x480 HTC Universal with a plastic keyboard felt antiquated compared to the 320x480 iPhone, especially starting with the 3GS

Lammy 1 year ago | | |

> The initial idea (2004-2005) was to build a Mac tablet computer based on touch screens.

There was lots of speculation about this starting in 2002 when “Inkwell” handwriting recognition showed up in Mac OS X Jagwire:

https://en.wikipedia.org/wiki/Inkwell_(Macintosh)

https://www.macworld.com/article/155597/wacom.html

https://www.oreilly.com/library/view/mac-os-x/0596004605/ch0...

And related patent filings go as far back as 2000: https://patents.google.com/patent/US7564995B1/en

cyberax 1 year ago | | |

> My 640x480 HTC Universal with a plastic keyboard felt antiquated compared to the 320x480 iPhone, especially starting with the 3GS

Opera Mobile existed at that time, though.

amarshall 1 year ago | | |

> Yes, you had to double tap on complex, non responsive websites

I suppose GP’s point is that the vast majority of websites in 2008 were “desktop”-only.

jbverschoor 1 year ago | | |

But the double tapping was intelligent to zoom exactly what you meant.

Also tapping links was very easy

wil421 1 year ago | |

When I saw the first iPhone commercial I knew it was the future immediately. Pinch to zoom and all the pleasures of a touch screen. Plus a web browser comparable to a desktop but with mobile usability.

The web experience and usability of the original windows phone and blackberries was terrible. Nokia had a Swiss Army knife that didn’t sell or translate well to the iPhone/Android future.

mikestew 1 year ago | |

320x480 screen meant normal web pages wouldn't display properly

I’m not sure what you mean by that. The whole point of the demo was it rendered the New York Times website just as it rendered in desktop Safari. In contrast to, say, Windows Mobile which would butcher the rendering. If the touch targets were too small, well, through the wonders of multitouch, pinch to zoom in.

the web was unusable because it required rewriting every page since existing pages were unusable

You’re remembering a much different iPhone than I am. Are you sure not confusing it with WinMo 5/6? Because then we’d be in some agreement.

bryanlarsen 1 year ago | | |

It rendered just fine on the sharp Zaurus or the Nokia phone sized Linux tablets, and was a much more pleasant experience than the iPhone pinch and zoom and pan and scroll.

But but on any mainstream phones. The phone I reference didn't exist. It could have and should have. If it did, perhaps Nokia would have survived.

Gigachad 1 year ago | |

Well to be fair, people did demand you be able to install apps and this feature was added shortly after.

I guess phones getting new features via updates was fairly uncommon at the time though.

ranger_danger 1 year ago | |

> Why would anybody buy an iPhone

Curious how you would respond to this argument today.

mparkms 1 year ago |

This is pretty similar to the Japanese "flick" keyboard that's fairly commonly used on smartphones. Instead of 3 possible directions per button there's 5 (up, left, right, down, and neutral): https://www.youtube.com/watch?v=Z5UEsHEZWII

It's pretty intuitive because Japanese kana is a syllabary that's organized by their starting consonant and one of five possible vowels in the Gojuon system: https://en.wikipedia.org/wiki/Goj%C5%ABon

I never got used to it but people who use it swear by it. Google even made a mechanical version for an April Fools a few years back: https://youtu.be/5LI1PysAlkU

kristopolous 1 year ago | |

I use it. It's great!

I'm native English but native Japanese are super fast. It's like watching a speed cuber

soraminazuki 1 year ago | |

It's 5 directions for Japanese kana, but 3 for the alphabet. Regarding usage, I find swipe input to be faster for English although flick input gives me more accuracy. In the end, nothing beats a physical keyboard though.

endofreach 1 year ago |

Wow, i always think "i wish there was a demo" of each iterative step on these kind of design journeys. And especially the iOS keyboard, as i thought a lot about it. Awesome that you made this.

I really like their idea to make the touch keyboard work well by increasing the (invisible) "padding" area for the key most likely to be typed next. So obvious in hindsight but demos like your's show part of the journey.

There are rumors that in the beginning they tried the ipod wheel as the user input interface.

While working on my device, that idea is super motivating. I know where i am headed and have done a lot of work and really got something very, very interesting already. But a few parts are yet not clear at all. But, i am definitely further than our "clickwheel" stage. Not yet at the "secret padding" stage for some input ideas, but still.

In case you are free and interested to loosely talk about ideas or feedback you have for a very novel device (and interface), please let me know how to reach out. Just because you made this demo, i feel we'd get a long & i can learn a lot from you. And this might be a very interesting challenge for someone like you. Anyway, great work!

juliendorra 1 year ago | |

Thanks! Yes, the blob keyboard opened the way to the idea of decoupling what the user see and what the software keyboard takes into account.

Feel free to contact me on LinkedIn! I'd be happy to chat

bramhaag 1 year ago |

Unexpected Keyboard [1] has a similar concept: you can tap keys to enter letters or swipe to the corners to input special characters. It also has all of the usual modifier keys so it's really nice to use together with Temux.

[1] https://github.com/Julow/Unexpected-Keyboard

collingreen 1 year ago |

This is cool! Thanks for building it and sharing it. I think phone keyboards are simultaneously amazing that they work at all and also still need huge improvements. There was a moment in the early mobile app explosion where we had some cool experimental keyboards but they mostly fizzled out, were acquired and shut down, or didn't reach their potential (looking at you, keymunk). I still think about this space all the time but, like password managers, it requires such a vast amount of trust I think it's a hard business to get into.

Was hoping the LLM boom would help us get sane autocorrect to help bridge the gap but so far that hasn't happened either.

juliendorra 1 year ago | |

Sometimes the best ideal solution is not the one that can win in practice!

The web was and still is a very bad hypertext system (no bidirectional links, no versioning, no integrated index, no transclusion, etc.) and we are still paying this debt! But this was also the an incredible hypertext system, permission-less, decentralized, and that made it a success.

I feel that the iPhone keyboard was the same, it had to appears (including in screenshots) innocuous and "without quality" to win over users.

werecat 1 year ago |

Neat prototype. While the key layout is unfamiliar, I could definitely get used to this. The layout reminds me of old cell phone typing, where each number had a set of letters associated with it and you had to press multiple times to get the letter you wanted. I wonder if testers at the time got confused trying to type like that.

Ntrails 1 year ago | |

Yeah i had the same vibes and could definitely see a refined version of that being very quick to type on

gfiorav 1 year ago |

Spent maybe 3 min with it and got the hang of it. I thought no qwerty was going to be a deal breaker but I think I could get "fluent" with it in a day or two.

The worst parts are: no upper case and the fact that there's no connection between what you typed and the keyboard once you submit.

juliendorra 1 year ago | |

Yeah, the goal of this particular prototype was to try and find a way to make typing letters actually possible! The iPhone team at this moment was in real trouble, as the naively 1-to-1 keyboard was absolutely unusable!

They stopped the whole 15 developers iPhone OS team and asked all developers to only build keyboard prototypes until they all would have something worth of a demo. As far as I know it's unheard of in the history of OS development!

Ken Kocienda was not in charge of the keyboard at all originally, he stumbled upon one working solution, and that's really what I found interesting in this prototype: it's a step in the iterative process.

dvdkon 1 year ago |

Thanks for making this! I hadn't even heard of this keyboard layout prototype until now.

I have a few friends that use gesture-based keyboards similar to this, and I myself use gestures to type diacritics and punctuation. So this idea is still alive after almost two decades, just not mainstream.

spiffytech 1 year ago |

Reminds me of Thumb-Key: https://github.com/dessalines/thumb-key

I'd be interested to try a full-featured Android keyboard like this.

jopicornell 1 year ago | |

I use Unexpected Keyboard on android and it has been amazing. It gives you so much options and speed that other keyboards aren't even remotely close, and seems very close to what you've linked.

cpeterso 1 year ago |

One of my favorite virtual keyboards was a linear swipe keyboard called something like “Minimum” (minus some vowels). The QWERTY keys were laid out horizontally like a piano keyboard. Text input was fast and surprisingly accurate.

This was about 15 years ago, but then the app and company disappeared. I can’t find any trace of the keyboard or company on the web today. Perhaps they were sunk by some patent issues.

Retr0id 1 year ago |

Fun, you can type "hello world this is a test" with only one swipe needed. "swype" keyboards could be seen as a continuation of this idea. At least, I think that's the chronology? Wikipedia says Swype Inc was founded in 2002, but the keyboard product didn't release until 2009.

juliendorra 1 year ago | |

The blob keyboard was never publicly shown (at least not until Kocienda's book in 2018) and was quickly morphed into something else internally.

So Swype probably reinvented the idea independently (as others did too I think)

Apple did patent some swipe behaviors for keyboards at the time of the first iPhone: https://patents.google.com/patent/US8059101B2/en

DerJacques 1 year ago |

This is really cool! What a neat little exploration.

I didn’t know about the history of the iOS keyboard, and this was a great insight.

juliendorra 1 year ago | |

Check the repo for more infos, and be sure to read at least the interview with Ken Kocienda, and maybe his book!

michaelhoney 1 year ago |

This is excellent and I appreciate your work. I love seeing prototype interface, and this one has real promise. I immediately found myself thinking about alternative layouts, particularly adaptive ones with common letter-pairs appearing radially.

grishka 1 year ago |

I honestly never understood why on earth QWERTY became the standard for touchscreen text input. It's a good keyboard layout but only when the keys are physical and large enough that you can type with two hands without looking at the thing. But it's absolutely godawful when you have to use it with one finger or two thumbs and no feel. All those dictionary-based crutches that were piled on top of it over time don't help, they only exacerbate the problem, especially considering how laughably poorly they work with languages whose grammar is more complex than English. I've been using touchscreen phones since 2011 yet I somehow still make typos all the time. This suggests that the problem isn't me.

An ideal touchscreen text input method should embrace the strengths of the medium. Touchscreens allow swipes and gestures, so use that to your advantage! Like in the OP, yes. I myself experimented with a similar concept, except it was a 3x4 grid of square buttons that each can input one of 5 characters, one for tap and the other four for each of the directions, inspired by Japanese keyboards. I tried to use it for a day, it worked, but it felt somewhat unnatural. Maybe I haven't given myself enough time to get used to it, or maybe it was a terrible idea to begin with, or maybe the character layout wasn't optimal (I arranged them alphabetically).

Another idea I have is to take the 8pen[1] concept but add more lines, make it 6 or 8, so it's practical for Russian which has 33 letters. Haven't tested this one yet.

I feel like touchscreen text input is a very under-researched area for a world where nearly everyone owns a smartphone.

[1] https://www.youtube.com/watch?v=q3OuCR0EpGo

kristopolous 1 year ago | |

Qwerty is the lowest cognitive lift.

Most tech people really don't understand how valuable "not learning shit" is to a product's success

grishka 1 year ago | | |

But there isn't that "not learning shit" aspect to it because none of your muscle memory translates between the full-size computer keyboard and the crammed mobile version.

And people somehow learned to type on phone keypads before. No one complained about it once they figured it out in the first hours after getting their first phone.

reaperducer 1 year ago |

Reminds me of the physical keyboard on my old SonyEricsson M600c.

Each key on the keyboard was a rocker switch. Push left for q. Push right for w, etc.

smitelli 1 year ago |

The thing I found interesting in trying to type out a single test sentence is how many of the letters were reachable with just a tap. It wasn’t until I really studied the layout closely that I noticed that it wasn’t in true alphabetical order. Oddly intuitive, although I would probably despise using it long-term.

juliendorra 1 year ago | |

Right, I noticed that too! (especially typing in english, might be different in other languages)

That's why it's totally different to learn about an UX and actually experiment it, we can notice these things.

But, all in all, I think you would probably get used to it long term!

The QWERTY layout was just a fluke, not the super optimized layout we retcon it to be.

Even its designer later designed different layouts! But the particular typewriter it was designed for had a big commercial success, so there was marketing convergence. It's of course wholly unadapted to typing on glass!

readthenotes1 1 year ago | |

The most common English letters are found at the top of the triad, and some only have punctuation as the other option -that makes for easy navigation

ano-ther 1 year ago |

Cool. And nicely implemented.

Having texted with T9 this didn’t feel too alien. Just the accept button slowing me down.

torben-friis 1 year ago |

Isn’t this basically just t9 from the physical keyboard era? I was very used to it as a teenager.

Wowfunhappy 1 year ago | |

With T9 you had to tap multiple times to get a secondary letter. Here you can tap or swipe to get any letter without multiple presses. I think that's a huge difference.

Izkata 1 year ago | | |

T9 was single tap per letter and an internal dictionary would figure out which letter you meant, then 0 would cycle between possible matching words, with the last option falling back to the more common multi-tap input you're thinking of (which I don't recall ever having its own name) to add to the dictionary.

ThrowawayTestr 1 year ago |

You should widen the margins for what's considered an up swipe

juliendorra 1 year ago | |

There's no up swipe in the prototype made in 2005 by Ken Kocienda (which this is a reproduction of). It's tap for center letter, and swipe for the sides.

What do you mean by "widen the margins"?

In any case, Ken Kocienda validated this reproduction's behavior as very close to his original blob keyboard (see bluesky posts between him and me)

The goal is not to make it better than the original prototype! But to the contrary, to give you the same experience that pushed Kocienda to explore further and design the right user experience.

vik0 1 year ago |

This is great! Thank you for sharing!

mutant 1 year ago |

You can spell “this is the shit” without needing angled letters