Agreed. I average almost 10 wpm slower on his test than others. The capitalizations, frequent quotations, and occasional phonetic spellings all make it a bit tougher than other tests (and real typing).
[bunnyfly]

Some other typing test to compare it to: (I consider these to be the best, along with Ryan Heise's test)

http://freetypinggame.net/free-typing-test.asp (freestyle typing–you can type anything)

http://www.keybr.com/ (you use your own input text; once you make an error, you cannot move on)

http://play.typeracer.com/ (goes by words, and you have to stop if you type the word wrong)

I can type fastest on keybr.com, probably because I don't have to ever hit backspace.

Hi, guys! I hope you don't mind this late reply :-)

Yesterday, I spent about 4 hours going through the 7,000+ sentences and removed all unnecessary quotes, fixed incorrect capitalisation and spelling, and removed sentences that contained non-English words. I also added some new options to the preferences:

- On screen WPM display
- Font size
- Double/Single space after sentences.
- Spaces or no spaces around M-dashes.

I've also posted a more decent highscore of 130 wpm. Go, colemak ;-)

Awesome changes!

My program doesn't really do anything out of the ordinary. It just receives keyboard events in the form of ASCII codes, and then renders them on screen. I can only conclude that your operating system is sending the wrong ASCII codes to my program.

This is probably quite likely. What is happening every time you type the letters 'th'? It does not come out as a real 'th'.

I see... Maybe you forgot to clear the modifier in your xmodmap file (see how capslock is handled for an example).

Ah, sorry, I missed the "+ 1". I tried a similar experiment:

keycode 109 =  Mode_switch   Mode_switch
keycode  10 =            1        exclam       exclam       exclam

109 is my right control key, but I configured it to play the role of Alt-Gr. Right-control + 1 will now type a !, and it works in the typing test, too.

I don't know how to fix your problem. I am guessing that you can temporary bind the ! symbol to another key combination that does work in the typing test. If you can do that, you could use xev to see if there is any visible difference in the kind of event that gets generated between the two different key combinations, and that might provide a clue.

I don't think there is any obvious way to change the typing test code itself. It is following the standard Java way of reading keyboard input:

class MyTypingTest extends Applet implements KeyListener
{
...
    public void keyTyped(KeyEvent event)
    {
        char c = event.getChar();
        drawCharOntoScreen(c);
    }
}

I have already suggested to you where you can start your investigation (look at the xev output).

Certainly, my code is correct (I've even shown it to you in case you want to dispute it). If there is a problem, it must lie elsewhere, such as with your xmodmap file or with the particular Java Virtual Machine you're using. If it is a problem with your xmodmap file, it is simply a matter of you trying different settings and reading the manpage. Otherwise, if it is a problem with your Java Virtual Machine, you could try a different one from a different vendor (on Linux, you have the choice of either Sun or IcedTea), and if that fails, you can report a bug to them.

The help I can give you is as follows. My understanding of the way that the Java Virtual Machine (JVM) works is that keyTyped events occur only for keys that cause a printable character to pop out. A modifier key on its own never generates a keyTyped event. None of this is under my control, of course, it is just the way Java was defined. It is certainly a possibility that the people who developed your JVM "hardcoded" the set of key codes for what they assumed were modifier keys rather than getting this dynamically from the operating system environment. If that is the case, then it is certainly a bug in your JVM which you could report to your Java vendor. Another possibility is that the xmodmap file is not correct and is causing the JVM to become confused about whether a particular key is a modifier or not. If that is the case, then you could try to fix it by experimenting with your xmodmap file.

If you don't want to file a bug report, and you don't want to fix a potential problem with your xmodmap file (if one exists), then the best alternative is to remap your layout so that the ! symbol can be typed without Alt-Gr.

But in any case, this is definitely not something that should be fixed in the typing test itself.

I really didn't like this comment. Are you laying the responsibility on me for "choosing" not to fix the problem? What exactly is "up to me"?

That is not fair, because as I have tried to tell you before, the problem lies in your environment (it could be the JVM/SCIM/Xorg or some other layer of system software that has to do with key codes).

The BEST that I can do in this situation is not to fix the problem, but to provide you with facts which might help you to tweak your environment in the right way. What saddens me is that I have already spent about 3 hours researching all of the advice I gave you above, even though it was not my responsibility to, and you are not interested in it. You are still hoping for an outcome that involves me fixing a problem, and that is probably why you are still insisting that there could be a way to fix it in my code. I have shown you my code, and you can see for yourself that the problem does not lie there.

I'd like to say, finally, that this is a hobby for me. If I were an actual company, you might expect me to go a whole lot further to fix your problem simply because it would equate to having more customers. But I am not a company, and this is a hobby. The typing test was not even designed for any locale besides the English locale because, after all, the sentences in the typing test are in the language of English.

One final thought is that you can try to change your locale to English before running this test. Maybe it will convince the environment to behave the correct way. That is:

LC_CTYPE="en_US.UTF-8" firefox

Hopefully the balance is right, now. Quotes will not be eliminated completely, because a good typing test should rate the well-rounded typist.

The quote symbol is certainly not difficult to type. It is a matter recognising patterns and learning the most efficient way to type those patterns. In these sentences, for example, it is usually the case that the letter immediately following the quotation symbol is a capital letter (possibly separated by a space), meaning that you can hold shift for that entire sequence.

Thanks for your efforts with this app, Ryan. It's one of the better-behaved applets I've come across, as a matter of fact. Several other web-based tests stumble when you use the keys normally designed as VKEY_# keys (in Windows at least) or use a program to insert your input into the input stream (like PKL does); yours manages that just fine.

(On a side note, input stream insertion might be a way of cheating in a typing game. It's about the lamest thing conceivable, but I have heard of such cheating. Not sure what could be done to prevent it...)

I don't have any AltGr chars left in my own layout that would come up in your test. For what it's worth, I made a PKL .ini file that uses AltGr+o for comma (and the comma key for ó) and that works just fine. But it's not worth much I know.  :)

That's 90% intentional ;-) When I made Tessellate (another game on that site), I designed it so that users themselves could enter the game rules into the URL and in effect design their own game. In the latest typing test, I added the same kind of URL support, but didn't advertise it. Unfortunately I forgot to change the description for custom-made tests, so for the time being it will still say 2 minutes.