Tuesday, January 31, 2012

Not learning steno: Day Five

Sorry - deadlines.

I was struck by the fact that stenography makes transcription a lot more lucrative than QWERTY typing. Here are a brief list of links to transcription job services I found with fifteen minutes' research a couple of days ago.

Sunday, January 29, 2012

Learning steno: Day Three

Still just chugging along in blind letter mode on Fly. I do a short warmup with the input chord and key highlighting displayed, then turn it off and see how many chords I can remember without looking. I'm getting better.

Interesting article in Wired on learning: recall is the best reinforcement, but it's most effective if you just barely recall the thing you're learning. Practice again too soon, and you're not learning as effectively as if you've nearly forgotten everything. I'm running Fly for ten minutes or so, twice a day, and I think it's probably about as effective as it could get.

Saturday, January 28, 2012

Fly: chord memorization drill

I think the first thing I need to do is chord memorization. I'm finding that if the chord is displayed anywhere on the screen, I simply type the chord, without actually really thinking about the letter it represents - this is bad.

So what I can do is turn off the input chord display and letter highlighting; I keep the steno key display on because I'm still memorizing the key positions (although I have to admit I've nearly got them down after only about fifteen minutes' total practice). I can then re-enable input chord display when I get stuck, which at this point is essentially for any chord that's more than one letter, except Q = KW, which is easy.

The problem with this is I have to take my hands off the keyboard and click the mouse. It would be way better to hit a key combination and have a chord hint displayed for half a second, after which it would automatically disappear again. That shouldn't be too hard.

Friday, January 27, 2012

Learning steno, Day One

Right now I'm just going through the Fly letter mode to get a feel for where the keys are. I did about ten minutes earlier, and my goal is to spend maybe half an hour each day, which I will deduct from my Reddit time, so no freaking harm done.

As a note - I thought I was a fast typist. In fact, I know I'm a fast typist. I'm just ... not as fast as I thought. 75-85 wpm. In fact:
Typing Test ScoreTyping Test Score
There's the two badges I just earned at the first response on Google if you want to try it as well. Let's just say my typical typing speed, when focused, is about 80 wpm.

Here's the thing. My income as a technical translator depends almost entirely on my typing speed. I translate about 4000 words per nominal day. At 80 wpm, that's 50 minutes of typing, but of course I can't actually maintain that speed (I felt as though my speed on the test was faster than usual, more fluid - and I spend a lot of time correcting mistakes, which I didn't bother with on the test; steno is reportedly more accurate due to its built-in understanding that you're typing English). 4000 words takes me four to five hours, mostly due to fatigue and boredom; depending on the source text, surprisingly little of that time is spent looking up terminology.

My feeling (and my hope) is that faster typing will increase my throughput and comfort, allowing me to handle more words per day comfortably. Because if it's day in and day out, comfort is paramount.

The only thing I've done during my translation career to improve throughput is a typing accelerator, which I use for words I see often. What "often" means varies. I hope to replace the typing accelerator with Plover, actually (which means I'll be subverting Plover to a certain extent), and we'll see just how difficult that ends up being. But when I started to use the typing accelerator, my output increased sharply, by about 20%. Hence my great hope for steno.

So, Day One. 80 wpm on qwerty, and that's my goal for steno, as quickly as possible.

Fly on Windows working

Plover itself (which is embedded in Fly) uses a thread to listen for keyboard events. This is entirely appropriate for Plover, but not good for Pygame, so I disabled it; I think we're going to want to redesign Plover/Windows a bit to move that thread out where it won't bother us.

But this marks the day I start learning steno. Note: it's freaking hard. I might want to tweak Fly a little; I'm not sure I like its lesson setup - but I think that's doable.

Monday, January 23, 2012

Fly on Windows: debugging Pygame/SDL

I've been spending a little too much time mystified by Pygame's inexplicable refusal to see any SDL events under Windows. Today, I discovered an incompatibility with the central StoppableThread architecture of the application, so I think maybe I'm on the track to getting this thing running.

Do you have any idea how long it's been since I tried to debug somebody else's code? In a language I'm rusty with, and using libraries I've only read about? Makes me feel like a teenager again.

Anyway, I just wanted to count coup here on the blog.

Update: Indeed, here's a StackOverflow thread describing substantially the same problem under OSX. There, too, the code seemed fine under Linux, and no events fired on OSX. Moral of the story seems to be: don't put Pygame in a thread.

Update 2: I put all the commented-out code back and sure enough, the window gets its messages - and then if I enable Plover I see the same behavior. So it's going to be off into Ploverland to see what's going on there. Weirdness on all sides.

Friday, January 20, 2012

Typing English

Actually, a certain amount of stenography's power is monitoring the typing stream and fixing things up based on the assumption that you're typing English.

Think about that a minute.

If I could do that all the time (well, when I'm typing English), then a lot of typing errors could be autocorrected as I type. I could also already put in some shortcuts - non-steno shortcuts of some kind - that could examine the text stream and act correctly, pluralizing words, adding -ly, that kind of thing.

And another thing - if I'm replacing AHK with Plover in a QWERTY mode, then I also get to define my own stop keys. That's going to be pretty damned convenient when using CAT tools! No more funky mistypes at the start of translation segments.

This is gonna be so cool!