Hi.

I feel your pain, my brother. However then I lose an entire word's worth of statistics and the statistics is one of things I'm most concerned about. Without some kind of key to "start the clock", I don't really know how fast you typed the first character, nor can I know how fast you typed the first word, nor the first trigram, so everything has to be thrown away there, which I felt "sullied" the stats-collecting routine. :( And I am too anal to just wing it by putting in a fake number. It is my sincere hope that the user will get used to it and incorporate it into the "flow" as I have. :) I could make the key customizable though if enter or something would be better?

Yes, sure. I'd say that is how you turn off the repetitions. :) I mean I could make a separate checkbox somewhere, but in my mind it would be redundant.

I know it's lacking some in user friendliness and intuitiveness... I guess I can make 0 for the boxes the default.

You can just find it in the list under sources, select it and click toggle disable (with the regex box in its default empty state -- which will match everything anyway). If there ought to be an even easier way than that then I don't know what you mean.

Yeah, I know.  I forgot to care about stuff like that before I just rushed it out. Sometimes I refer to texts as lessons and sometimes as texts, too. (However they are meant to be separate in the database. Lessons are only those texts generated by the lesson generator. Or fragments, whatever. I guess I need some help documents.)

OK, OK, I made it optional.

But first off I think your "typical example" is a bit of a generalization although I caused that myself I guess: when I said I felt your pain I was more showing empathy than I was actually being in the same situation, because for me it has never been a quirk or an annoyance at all. I use this program like Typeracer.com. I don't seek to write serial texts at a leisurely, continuous rate, (I added the serial text feature fos a friend) I seek to race and go as fast as I can go, and at first I thought nothing of the space, it was something that followed naturally from the implementation, but later I have actually grown to like it as a sort of green light signal for my fingers. A dip before the sprint. (And space is a lot better than waiting 10 seconds on Typracer, which IS annoying.) I have never had anyone else comment on it either, though I guess you are like the third person to try this program, so 1 out of 3 is a pretty large chunk. << edit: Actually, strike that. I asked another guy who had tried it and he said he thought the space was odd, though he never voiced it before. :p Can't trust friends, it seems. >>

If you turn off the option it sets the time used for the first character to the time used the last time you were timed using that key. This should be natural enouth... (If this is the very first time you type the key, another approximation based on your average key-striking-pace is used.) As you suggest, this inaccuracy should get lost in the mountains of data to come and probably no one but me will be bothered by it. Problem is I am the wrong sort of idealist. The way my brain is wired I tend to choose the ideal of the numbers/mathematics of a programming issue instead of thinking about the lovable squishy meat popsicles commonly referred to as users. However I don't lose anything by making it optional and it wasn't too much of a hassle so that's what I've done.

I don't intercept shift keys since in the program I simply subclass a text box and track changes to it -- and shift doesn't do any changes. This is a conscious choice. I don't want to track shift, control, compose, altgr, etc. as then I would either have to mess with character composition or let go of keyboard agnosticism (what is shifted in one layout is composed on the next and a plain key on the third, etc). Second, I only want to care about the "end result". I mean maybe you have to press 3 physical keys to produce a "[" I don't want to have do deal with that, I just want to measure how fast you actually produce "["s.

GG reference? :)

A redesign of the source tab has been at least several days coming! :) When that happens I'll probably do something wildly different, I don't know yet. I'll have to let my subconsciousness work on it. But I've been thinking of a sort of bastard union of Performance and Source where there is a "root" node and the sources underneath, and you get statistics and graphs when you click on the various items in the tree. Having several thousand items like that in a tree is quite slow in Python though, so I have to figure out how to make it faster.

I do like regexes though, as they give unprecedented power to the users who knows them -- and in an ideal world (mine) everyone knows them! Maybe make a separate search/edit tab for finding and editing texts/lessons, I don't know. All I have are ideas.

Nope I haven't seen it. That's interesting, but doesn't give you word-stats, which is primarily what I wanted. I've sensed that there are a lot of left-handed-ish words (like "was" "wharf" "quaff" "war" "raster" etc) that kill me on Colemak when I try to go fast (I make the most mistakes and I am slowest with my left hand) and I wanted to identify them so I could train myself. However, a general key-logger that gives typing stats for all user programs is not something I have planned as it differs from Xlib to Windows to OSX to ...

<< edit: forgot to say, I uploaded it here this time: https://code.google.com/p/amphetype/downloads/list >>

No, thank you. Makes me feel all kinds of warm and fuzzy.

Yes, I've had this idea myself. Might make it an option at sometime. As well as make it an option to display "live" statistics of your current session, though I guess this can be distracting if made too blingy.

Anyway, you guys are free to suggest features at https://code.google.com/p/amphetype/issues/entry -- this way I have an easy overview and won't have to rely on my memory :)

Yeah.

wpm = 12.0*chars/secs

For the word/key/trigram stats I store the actual timings (otherwise I couldn't calculate averages on them in a sane manner), but display them as WPM cause it's easier for people to relate to that than "113ms", even though it doesn't make much sense that I type "n" at 150wpm. ;)

Added a couple of ideas tristesse. The first one is me but the missus was signed in to Google.

Thanks. Now it almost look like a real project. :)

Qt runs on OS X so the program should run on OS X... As long as you can install Python (2.5) PyQt4 (Qt 4.3+?) and matplotlib (though this won't be necessary for the next version I'm planning), then you should be able to run it just fine? But I have never touched an OS X and unless I get access to one in the near future (unlikely) or someone else figures it out... then I dunno.

I mean, in principle I don't think there's anything stopping it from running on OS X, it's just that I know nothing about distributing to OS X.

Like last time I tried the sources on my Linux laptop that I left running back home (visiting my dad for christmas) it worked just fine.

oh my god!    GG references!

Would it be possible for you to distribute Amphetype as a zip archive in complement to (or instead of) the rar archive? It would make it easier for most people to try out the program, I think.

Do you have to get a Google account for that to work?

I tried the no-Space-to-start approach and it worked great for me... until I reached the end of the page. It didn't proceed to the next page? This happened when I first unchecked the option and then imported and started a text. If I started typing a text with the option checked and then after the first page unchecked the option, it worked as it should and proceeded to the next page automatically, but then I ran into the same problem again after a while.

My Amphetype.exe.log file contained a couple hundred instances of this message:

The text pages are only a few lines long. Is there a way to make them as long as I want? I tried increasing the "try to limit texts and lessons..." option, and then re-importing a book, but this didn't seem to make a difference to the "lesson" lengths. I had done a few "lessons" on that book, and the statistics still showed when the book was deleted and then re-imported; maybe that is of significance?

Could you make the program count a single space as enough when the text contains a double space, or convert multiple spaces in the text to a single space? I ran afoul of a text that used double space after a period, which isn't typographically correct (I think?!); it took me a while and some cursing to figure out that I had to type a double space to proceed where I expected that a single space would be enough. That isn't your program's fault but the stupid text's, but if your program could smooth out that road bump it'd be nice.

The text box turns from white to black when an error is typed. That shocks me every time.  :)  Not sure whether a shock is beneficial to my progress or not, but maybe there should be a way to customize this? I'd like to be able to pick background and text colors, and then the background and text color for when mistyping occurs. That way I could for instance turn the text from blue to red while leaving the background color unchanged whenever an error was typed ("hi-games style"), if I preferred that. For me it would be quite acceptable to enter the colors as for instance #rrggbb values in input boxes, but I suppose that color picker modules are also readily available.

Maybe there could be a way to browse lessons while in the typer mode? Let's say I'm typing a book and want to skip a chapter or go back a chapter. I'd like arrow buttons at the top or something. If the chosen selection method is Random there could be only one button instead, maybe with an asterisk on it to signify refreshing/reselecting.

This looks like fun, but it's a .exe... is it available for Mac?

Smee again.

SpeedMorph & keyboard samurai -- someone contributed an OS X build on the project page, you can see if it works. :)

You don't know me either! ;) But glad it works.

Which one? (And specifically, what encoding did you chose when you downloaded it..) Did it give an error message (or was there a *.log file in the directory after you closed the program)?

Hmm, yes, that's probably annoying. For now you only have two options:
a) You could find the *.db file in the amphetype directory and delete it, that clears ALL your stats and starts fresh. It will have the name of your username on the system probably.
b) Download sqlite3 client and open the .db file and remove whatever you fant manually from the database tables (in this case the "result" table)

Occasionally there are some non-printable characters in the text that show up as [square]. When such a character occurs, it is nigh impossible to proceed with that piece of text. It might be prudent to filter() the text prior to it getting to the user.