dvorak layout

[bhtooefr]

Man, this is hard.

Both accuracy and speed are pretty crap right now. End up having to switch back to QWERTY now and then just to rest my brain, and actually get stuff done. I know, that is bad.

Edit: Discontinuing the experiment. I know, that's giving up easily. Not because I'm having trouble with Dvorak (I'm starting to get the home row), but rather because Dvorak requires conventional use of the home row, and my typing style on a QWERTY board (which moves the hand a lot more) appears to actually avoid hand strain caused by the left pinky being overstressed typing on As (I generally shift one key out from normal home row positions, and just move my hand as needed to hit the keys I need to hit).

[Vierax]

bhtooefr > Do you learn dactylo typing on your new layouts ? It's hard at the beginning but it's worthy.

[bhtooefr]

What is "dactylo" typing? (In English, "dactylo" is a prefix referring to the fingers (and sometimes toes), but I'm guessing it has to do with a certain method of using one's fingers on the keyboard?)

I usually use my ring finger where the "proper" finger is the pinky, on QWERTY. Pinky is used on the extreme right and left keys (shift, caps lock, tab, Ctrl, enter, backslash, backspace, that sort of thing.)

[Vierax]

I mean Touch Typing

[fossala]

IMO touch typing is more important than learning a new layout.

[Vierax]

IMO touch typing is more important than learning a new layout.

We are agreed Moreover, it allows you to be independent from any marked layout and have a blank keyboard if you want to.

[bhtooefr]

I'm exclusively a touch typist. (I don't do it properly, but I do do it - don't ever look at the keyboard.)

[Input Nirvana]

I didn't completely comprehend the benefits of touch typing until I learned. I understood in principle, but it was dramatic once in action.

[fossala]

Just thought I would let you guys know, I'm back to dvorak. Got to around 30wpm in maltron but It's too much of a commitment.

[Input Nirvana]

Fossala:
I sent you a PM, but it is hung up in my outbox along with another message I sent to someone a couple days ago. I tried logging out/in and such stuff. ???

I'll send pics and follow up and you can decide

[fossala]

I got the pm, thanks.

[Jim66]

I was all set to start learning colemak this weekend but I can't seem to get it installed in windows 8. It's pretty strange that they have dvorak as standard but not colemak.

[bhtooefr]

Windows has had Dvorak available for ages, they haven't adopted Colemak yet.

[Icarium]

I'm using autohotkey and it works well enough for me.

[bpiphany]

I never got around to actually change the key mapping, I just redid the physical key layout for my symmetric stagger. The plan was to come up with a layout of my own to fit the type of typing I actually do myself.. A key logger and some number crunching should do that trick, right? =)

I'm able to create my own layouts in Linux. It's very messy, but doable. I have never even tried on Windows, but has anyone had a look at this? http://msdn.microsoft.com/en-us/goglobal/bb964665.aspx

[Hubbert]

SUKA:
There is still some work that can be done on keyboard layouts, but it may be safe to say that the 'big' jump towards improvement has already been made and future layout improvement will probably be relatively minor. At some point it will be a zero net gain.

This...human anatomy hasn't changed for a while, and there are always tradeoffs. I am intrigued by CarpalX beating Dvorak by exchanging only 5 pairs characters (I'd like to confirm his research myself before taking the plunge).

To further Webwit's point, though, in addition to not having computers to do analysis some of Dvorak's false premises (such as hand alteration) may have been more valid with heavy mechanical typewriters than with keyboards.

More than the alphabet, I'm looking at (and may make a separate thread about) the arrangement of punctuation. I'm thinking of making another layer for the hard-to-reach punctuation keys...but some hardcore typists claim that modifiers slow them down greatly.

[Soarer]

... human anatomy hasn't changed for a while ...

Well... Americans Getting Taller, Bigger, Fatter, Says CDC

[Hubbert]

Yep...but sadly the UK is not far behind. Mexico and Quatar compete for #1, though.

Back to keyboards, if you take CarpalX analysis at face value, his 3 swaps are < 10% worse than Dvorak and 4 swaps is more than 5% better. Even then I start to wonder how it would affect shortcuts in Emacs (many of which I have remapped).

I'm thinking about analyzing myself in a typing program that does per-letter and per-finger analyses, which would help me tune CarpalX. One issue I am finding with that program is that he assumes textbook typing patterns, and I don't really think that right pinkie should hit -, for instance (especially not straight men and lesbians, who have longer ring fingers).

[dirge]

Longer ring fingers? Eh?

[Soarer]

Heh, the statistic I actually started looking for was how much taller people are on average now, compared to when typewriters first used 3/4 inch key spacing... it must be a few inches. Anyway, we are changing

But yeah, if all we did was type text, it would be much simpler! As a programmer, probably about half the text in my code result from auto-completion, leaving punctuation as a fairly major part of what I actually input.

(I can touch type, but I usually don't.. at least, not 'by the book').

[Jim66]

Well I tired reallly hard to stick to DVORAK for 2 days but it was too difficult. It's just the most frustrating thing in the world knowing that you can blast a paragraph out in 2 mins and with a new layout it's going to take you 20 mins.

I don't think there is ever going to be a great time for me to switch .

[fossala]

I keep thinking about trying a new layout.

[Pby]

Well I tired reallly hard to stick to DVORAK for 2 days but it was too difficult. It's just the most frustrating thing in the world knowing that you can blast a paragraph out in 2 mins and with a new layout it's going to take you 20 mins.

I don't think there is ever going to be a great time for me to switch .

It takes time, two days is far too little. I remember when I first started with Dvorak it took a good two weeks to get comfortable with it. It probably took another two three weeks before I could type as fast as I could on qwerty. But I haven't looked back since switching and its almost ten years since I've switched.

[Hubbert]

Longer ring fingers? Eh?

The theory was that the ratio of index finger to ring finger length was correlated with prenatal androgen exposure and a variety of physical and behavioral traits. Googling "finger ratios and testosterone level" shows some studies from just this year refute what was a very common theory just a few years ago.

http://en.wikipedia.org/wiki/Digit_ratio
Note that it's the length of the fingers from joint to tip: you can't just make a Michigan[1] and compare Alpena and Traverse City because of the the angle of the palm (corresponding to the the Toledo Strip[2]).


[1] And yes, Michiganders really do use their hands as maps in conversations
[2] http://en.wikipedia.org/wiki/Toledo_War

[Hubbert]

I keep thinking about trying a new layout.

Another look at the CarpalX numbers:
QWERTY: 3
Dvorak: 2.098
CarpalX_3: 2.1 (make 3 swaps)

Dvorak improvement: 0.902
CarpalX_3 improvement : 0.900
CarpalX_3 improvement / Dvorak improvement: 99.778%

Even then, I'm still reluctant. I care more about mental overhead than ultimate typing speed; the last thing I want is some corner of my brain mentally translating stuff. I need more hard, personalized facts before making the jump.

[Hubbert]

Heh, the statistic I actually started looking for was how much taller people are on average now, compared to when typewriters first used 3/4 inch key spacing... it must be a few inches. Anyway, we are changing

Well, that helps: I think the standard spacing was too large, if anything.

But yeah, if all we did was type text, it would be much simpler! As a programmer, probably about half the text in my code result from auto-completion, leaving punctuation as a fairly major part of what I actually input.

And on topic for this thread about Dvorak, when I did post code samples into http://www.patorjk.com/keyboard-layout-analyzer/ , the results for Dvorak were pretty bad; it put even more load on the right pinkie than QWERTY.

[Icarium]

That is a known problem with Dvorak. Or at least a common complaint I hear from people who switched.

[bjarven]

I never got around to actually change the key mapping, I just redid the physical key layout for my symmetric stagger. The plan was to come up with a layout of my own to fit the type of typing I actually do myself.. A key logger and some number crunching should do that trick, right? =)

I'm able to create my own layouts in Linux. It's very messy, but doable. I have never even tried on Windows, but has anyone had a look at this? http://msdn.microsoft.com/en-us/goglobal/bb964665.aspx

I use it to create a swedish version of colemak, it's very useful. And then you get an installable you can take with you to any windows computer. My colleagues have started noticing that a "crazy layout" appears whenever they hit Alt + shift on some of the computers we have in common... the next step is to make them appreciate it, but that might take a while

[Soarer]

Heh, the statistic I actually started looking for was how much taller people are on average now, compared to when typewriters first used 3/4 inch key spacing... it must be a few inches. Anyway, we are changing

Well, that helps: I think the standard spacing was too large, if anything.

But yeah, if all we did was type text, it would be much simpler! As a programmer, probably about half the text in my code result from auto-completion, leaving punctuation as a fairly major part of what I actually input.

And on topic for this thread about Dvorak, when I did post code samples into http://www.patorjk.com/keyboard-layout-analyzer/ , the results for Dvorak were pretty bad; it put even more load on the right pinkie than QWERTY.

OK, so I started with a flippant point, but it - and the punctuation problem - are related in a sense: different people will find certain keys easier to reach than others. That means that scoring layouts without allowing for that can only be so accurate - a difference of 5% probably means little in a practical scenario. One part of the scoring is usually to 'dislike' the bottom row, isn't it? I actually find the centre and edge keys (just inside each shift) of it perfectly comfortable to use; certainly easier than reaching either ring or pinky up to '-'.

I mean to experiment with using a Fn-layer to move number row and right pinky punctuation into the home area (about 3/4 of them are accessed using shift anyway). Probably in a logical way, so that muscle memory isn't totally void, so e.g. Shift + 1,2,3... would be replaced with Fn + q,w,e... or Fn + a,s,d... (for qwerty, but of course the same move could be done on dvorak or colemak). It's harder to be logical about where the right pinky punctuation goes, but in that case the whole reason to move is to reduce strain on the pinky, so muscle memory isn't important. It would be interesting to see how adding such moves would affect the relative scoring of qwerty/dvorak/colemak etc.

[Hubbert]

Soarer, you read my mind! I've been thinking almost the exact same thing. I have a layer for the navigation, backspace, and escape accessed underneath the spacebar (pssst: that's why I asked if your protocol converter can also read one or two keys).

And I agree completely about the bottom row vs the "penalty box" with all the symbols: I'm tall and have long fingers and just get there: while I'm sure some would recommend floating the hands up there, but I basically conclude that it is simply un-touch-typeable: they are not really accessible from the home row, and there's no real anchor point either, unlike moving to the numpad or arrows...though I could think of creating one, what's the point of repositioning the hand for what is usually a single key). Now I can sometimes hit underscore (which is a shift but used far more than hyphen/minus (http://mtgap.bilfo.com/theory-of-letter-frequency.html), but often not.

I also looked at the Knight keyboards (and Sun to a lesser degree): they tended to have a lot of extra unshifted duplicate keys (i.e., parens on 90 but also on separate keys, a colon key in the upper left with the US-ASCII `~): this brings up the question of what the cost of shifting is (Sean Wrona says he uses CapsLock because it's easier than synchronizing shift, Xah Lee is also moving away from chording).

But really, there are 32 symbols (22 shifted, including the most common ones) on a standard US keyboard, so we'd need an entire extra keyboard to avoid shift, impractical even at a desktop. So if we are going to have to use a modifier, we might as well make the target keys closer to the home row in the process.

As far as keys to include, I'd probably omit the reachable numbers (and maybe punctuation): they really aren't used all that often, desktop keyboards have a numpad for heavy numeric entry anyway.

So I think that we need a test that truly measures the cost of each key position (i.e., scan code), and independently determines the shift penalty. The shift penalty could be determined by randomly capitalizing text (it can't just be at the beginning of sentences) throughout exercises (whether it varies by row is an open question). The shift penalty is probably not all that critical having made the decision to use a layer instead of adding multiple extra rows. The other advantage to an experimentally derived cost is that it accounts for different techniques, i.e., the implications of using the ring finger where theory recommends a pinkie, etc.). We would need to look at speed and also the amount of backspacing and look at the key in question but also the keys preceding and following the difficult key.

Then we would have true cost data to plug into CarpalX or a similar program. The 47 white keys on the US layout lead to 2.6x10^59 permutations. Time for a Group Buy with 1-877-CRAY-INC? ("Uh, and our organization has a unique requirement: we'll need a few racks of keyboard connectors...yes, we realize that it may incur some nonrecurring engineering...yes, Keyboard Interface Processor is an acceptable name for that feature.").
Another useful phone number: the ticket office for the Cleveland Indians baseball team: 1-216-420-HITS.

I don't think preserving same finger is important for muscle memory, since a key motivation of this exercise is that we probably don't have it. I was thinking of mapping something resembling the standard positions with normally shifted symbols on the home row and unshifted above it. The problem is that preserving same-hand is helpful; I disliked what Kinesis did with += (but that also had to do with the fact that the key involved was unreachable from the home row). And I would like parens on jk because they're common. I obviously have not progressed very far in my layout.

I also don't know if it makes sense to include the entire number row: 234 and 789 are pretty easy to reach already, and the associated symbols are less frequently used. But it might be cleaner.

(and I'm less likely to try any CarpalX alphabet swaps right now: E/K means I can no longer type hex with only the left hand, other swaps seem likely to lead to confused thinking, and my phone, GPS, etc., all use QWERTY. I care more about the qualitative change from "un-typeable" to typeable than the marginal efficiency gains from a better alphabet layout. It's like average driving speed: a 10 mph slowdown hurts your average far more than driving 10 mph over your target helps you. And also, losing an extra layer degrades gracefully, like doing without a numpad or mouse, rather than failing awkwardly.)