You can now create custom keyboard layouts and compare them against other layouts in the Keyboard Layout Analyzer. Below is a screen capture of the tab interface for the new feature.

Custom Layout Creator

Layouts 1-5 can now be configured any way you want them. So you can create your own layout and see how it stacks up against the more well known layouts, or you can make modifications to the existing ones and see how well your changes do.

Hopefully the interface is pretty straight forward. Clicking keys causes them to change places, and you can remember layouts you created with the use of the “Layout ID”.

I’ll be wrapping up my updates for this app pretty soon. I’m not sure what my next project will be, but I have a couple of ideas I’m considering.

I’ve added a new feature into the Keyboard Layout Analyzer that will let you know the distance your fingers travel while racing across different keyboard layouts. It’s interesting not only because it lets you know how far your fingers are crawling, but because it lets you test out one of Dvorak’s claims on why his layout was better than the QWERTY layout:

Dvorak estimated that the fingers of an average typist in his day traveled between 12 and 20 miles on a qwerty keyboard; the same text on a Dvorak keyboard would require only about one mile of travel.

Sources: [1] [2] [3] [4]

Below I have the results of several pieces of text I analyzed. The QWERTY, Dvorak, and Colemak layouts are compared against each other. The distance shown for each is just an estimate, but it appears as if the idea of typing 20 miles in a day is a little excessive. Also, the Dvorak layout only seems to give you a 40%-45% improvement over the QWERTY layout when it comes to the distance you type.

Keyboard Monument
(It uses a QWERTY layout)
Photo By jasoneppink (cc)

George Orwell’s 1984 book, Chapter 1:

QWERTY – 0.5629 miles
Dvorak (Simplified) – 0.3107 miles
Colemak – 0.2935 miles

Christopher Nolan’s The Dark Knight Movie Script:

QWERTY – 3.0929 miles
Dvorak (Simplified) – 1.8838 miles
Colemak – 1.8497 miles

Jared Diamond’s The Curse of Qwerty article:

QWERTY – 0.3852 miles
Dvorak (Simplified) – 0.2173 miles
Colemak – 0.1938 miles

Steve Yegge’s Pinocchio Problem blog post:

QWERTY 0.7167 miles
Dvorak (Simplified) 0.4115 miles
Colemak 0.3619 miles

My Previous Post:

QWERTY – 0.0412 miles
Dvorak (Simplified) – 0.0233 miles
Colemak – 0.0211 miles

The algorithm for determining the distances is still being tweaked at the moment, so the distances may change in the future, however, it wont be dramatic. The biggest thing I have left to do is to factor in the pressing of the “shift” keys when an upper case letter is typed.

"c" is the distance between the "F" and "R" keys

To the left you can see a diagram showing how the distance between two keys is calculated. The center coordinate of each key is known, so the Pythagorean theorem (c² = a² + b²) is used to calculate the distance between keys. The horizontal distance between the center of two letter keys that are next to each other is said to be 1.9 cm [reference]. This is 40 pixels in my set up. Therefore, I calculate the distance in pixels, and then convert it to meters by dividing by 40 and then multiplying by 0.019. I’m probably currently doing way more calculations than I should be doing, and that will be optimized later, but for now this is how its currently working.

If you’re skeptical about the distances you get back from the program, you can also try Colemack.com‘s Layout Comparison Tool, which also produces estimates on typing distances. They’ll be slightly different from what I display, but they’ll be within the same ballpark. As an aside, somehow I missed seeing that app until recently, but at least our programs have their divergences.

Let me know if you have any issues with the distance calculator. Also, I’ll hopefully have some more updates to the Keyboard Layout Analyzer coming within the next week or two.

I’ve uploaded a new version of my Keyboard Layout Analyzer. You can see a full screen capture of the output page below. The new version has a layout similar to the one I discussed in my last post, and has a different chart set up, as well as a hand full of other small new features.

The New Layout For Keyboard Layout Analyzer Output

This new version is essentially a complete re-write of the old version. I had been making mental notes of features I thought would make the program a lot more useful, and after thinking about how I’d implement these new features, I realized I’d be better off just re-writing the application instead of modifying the old one.

This re-write was also a way for me to “pay down” my “technical debt” (I think I like that term), as I was able to more efficiently implement certain things to make them a lot more flexible. When I write stuff for fun, a lot of the time I just go with the first or second implementation I think of and then get to work. This re-write will probably serve as a good lesson learned, since the re-write was a little more painful than I expected. Only one of the new features that I thought up is in the program right now (the new chart set up), I didn’t have time to get to the rest of them, though I hope to implement them in the next couple of weeks.

Here’s an overview of what’s in this version:

• Updated interface.
• Able to view charts of different layouts all at once.
• Able to select what pieces of information you want to factor into the charts.
• Addition of the Capewell and Arensito (Simplified) layouts. However, this set up isn’t final. These layouts will still be in the next version, but the options will be a little different.
• “Personal Layout” factored into the statistics.
• A layout recommendation is made based on what you typed.

This new implementation also consists of 26% fewer lines of code. 🙂

This new version was tested out in Firefox 3.0 and 3.5, Chrome, and IE8. Unfortunately, the new IE8 messed up my install of IE6Eolas, so I’m not sure how the app works in IE6 or IE7. Oh, and I hate to rehash the cliché of bashing IE, but IE8 is actually a lot slower than IE6 and IE7 when it comes to executing JavaScript. Enough so that I was actually a little dismayed. I had been hearing mostly positive things about IE8, but I just didn’t see anything that I thought was that impressive.

Anyway, if you have any suggestions for the Keyboard Layout Analyzer or find any bugs in it let me know.

Coming up with a good looking, easy-to-use design is hard. Sometimes I feel like I spend more time trying to design an application’s layout than I do programming its back end, and most of the time the resulting layout isn’t that impressive. I think I’m just better at creating functionality than I am at making something that looks pretty. However, no one’s going to look at something if it looks like crap, and people tend to give applications more of a fair shake if they look nice, so I think the time I spend trying to become a better designer is worth it.

Recently I’ve taken to trying to improve the look of my Typing Speed Test. Below you can see the initial design, a design I put up last month, and the most recent design, which I put up this past weekend.

Initial Design

Second Layout

Current Layout

I made the initial design with the idea of keeping things as simple as possible, however, in retrospect, the page looks kind of unbalanced and the big yellow explanation box at the top kind of comes off as a little bit of an eye sore. The second layout was made with the aim of being a bit more stylish. I got ride of the yellow explanation box, threw in some small gifs, added a keyboard reference, and put in a header bar to make it seem a little less plain.

The third, and current layout, was made with the aim of trying to make something that, in my mind, looked good. I don’t think I quiet accomplished that, but I think I came up with something better than what I had. The interface seems cleaner and more balanced, and I like the additional color at the top of the page. The only thing I’m not sure I like are the header images, they seem a little cheesy. So I’ll probably keep tweaking the layout. My goal is to come up with something I can re-use on future (and existing) applications.

Last time I said I had two more major features that I wanted to add into my Typing Speed Test. That number has since ballooned quiet bit and now I’m not really sure when I’ll be done with it.

Random Pro-FF Image

The biggest change you’ll notice today is that the input text is now in a 3-line format that scrolls once you’ve typed to the end of the second line. I like this change a lot sense I think it allows for a more simple user experience.

Another reason I like it is that it allowed me to find my first undocumented (for the most part) bug in IE (well, the first one that I’ve taken the time to examine). It appears IE has issues with returning the correct offsetTop position of span tags in certain situations. I’ve set up a test page here that’ll show you different offsetTop values when you use IE as opposed to other browsers. This guy had similar issues, though he couldn’t reproduce his problem via a test case. Luckily for me, his offsetTop span solution also worked for what I was doing (just set the span’s position property to relative).

My hot spot visualization after 1000 words of Moby Dick. I got kind of tired.

Since the input text can now be scrolled in a smaller area, I’ve also added in some additional test time options. You can now run the test for up to 5 minutes or up to 1000 words. I will warn you though, typing the 1000 word version of the test more than twice in a row will leave you kind of sore.

A lot of the other changes I made were under the hood changes. To decrease loading time, I’m now using the YUI Get Utility for certain script files. This utility allows developers to grab script and css files at run time and integrate them into their code. Right now I’m using it to grab unloaded script files that contain functions that return input text for the typing test. I figured that most users aren’t going to try every possible test input, so there’s no reason to load them all each time. It doesn’t improve loading time a whole lot right now, but it’ll keep the loading time from growing once I start adding lots of different test inputs.

I’ve made a bunch of other minor changes and fixes, but I wont bore you with all the little details (unless you want me to). If you find any problems in this new version please let me know.

I’ve updated my Typing Speed Test with a new feature that lets you see how fast you typed each key on the keyboard. Once you finish the test, you’re presented with a visualization of the keyboard where the keys that were typed faster are more red and the keys that were typed slower are less red. Keys that are completely white were not typed at all and grey keys aren’t factored into the calculations. To see stats about an individual key, all you have to do is mouse over the key you’re interested in. Below you can see a screen capture of the feature.

A screen capture of the Hot Spot Visualization feature. The mouse was hovering over the "W" key when this shot was taken.

Let me know if you see any problems. I still have two major features I want to add into this app before I move onto my next project.

EDIT March 26, 2009: I’ve updated the visualization. Blue now represents key that were typed slowly and red represents keys that were typed quickly.