• You are not logged in.

How about this layout??

  • Started by nawfal
  • 77 Replies:
  • Reputation: 0
  • Registered: 02-Nov-2010
  • Posts: 38

thanks kqr and spremino for your valuable inputs. i found them informative. Sorry for the late reply.

@ spremino, few things:
1. I didn't mean syntaxes never play a role at all. Rather I said, it is not what really is important. I was seconding what you have once mentioned.

2. Thats what, even arithmetic operations are in prefix style in lisp languages which is a turn off for a newbie. not for most. This was the point I was making.

3. And for rest of the detailed inputs, I'm grateful. I now understand why you have been picking OCaml as human time's something important. I'm reading the same thing everywhere. I would see to it as well apart from Python. I've to admit its languages like D and OCaml that seems more impressive for me apart from the easier ones like Python, Ruby or .net

Offline
  • 0
  • Reputation: 0
  • Registered: 02-Nov-2010
  • Posts: 38
SpeedMorph said:

This layout looks like one of the better ones I've seen and you seem to have really done your research, but I don't think it's as good as Colemak or the MTGAP-series.

Yup, technical specs-wise. But in my books, not by much. Moreover that depends on how big you people consider each hurdle really is. Same finger is something that's talked about the most, but i feel it's the home row jump thats hampers typing speed the most. Giving it twice the weightage of a same finger, I find my layout to be better, albeit slightly. It's not true for a standard horizontally aligned keyboards we confront daily. But I find it awful which almost limits the usage of pinky( keys under pinky doesnt fall under its natural position/movement). My ergonomic design looks like two rainbows under both ur fingers. Hope you get what pattern it form on a desk when you drum naturally.. And thats it.
         Secondly it can also be 'cos of the slightly different frequency data we have. Very much possible. Come on, we are talking about differences in 1/100th or 1/1000th decimal positions ! :)
         And lastly, the biggest advantage for both you and Shai, as to improving technical specs, is that you could be placing a punctuation under letter E. We know letter E is the most dangerous to handle, and same finger considerably drops when you have a punctuation under E. In my design where two "rainbows" are wide apart, a clubbed design with punctuations coming together with alphabets is not so pleasing for eyes. I personally feel it looks bad on a standard keyboard too, though not much in conventional Colemak like designs. I still feel I've made specs better (low) with this disadvantage of having to place an alphabet under E. ....... In short, what I'm saying is, the advantage you claim abt Colemak and Mtgap, if ever in reality, is also 'cos of same finger minimizing technique you have employed.
     If I'm sounding too pompous or you prefer concise replies I'm sorry. I hope you'll love such detailed discussions as well. I do.. Sorry for the late reply. I was busy.


SpeedMorph said:

The TH digraph is a unique solution and it has its advantages but I don't think it's very elegant. Touch typing works a certain way; why should you deviate from that just for this one digraph? I'd rather use a keyboard that keeps things simple..

I dunno how dis/non-elegant TH combination makes in my layout. I would love to hear what you have to say on this. There's hardly anything that disturbs the regular way of touch typing, except the fact that touch typing is made more efficient. Colemak, Mtgap (3.5 version), Asset and few other lesser known ones have placed H as in my layout. Moreover everyone outside the Workman school of thought wouldn't have a problem in having H where i placed it. We anyway have to move our index finger/wrist a bit to reach out to H. Now our middle finger would be on T (in my layout). Why wouldn't we utilize this undesired movement by pressing the T as well (in TH digraphs).
       It depends on how we touch type as well, is it more scientific just to extend our index finger to H position without moving wrist? or is it desirable to hit H by moving wrist so that middle finger fall on T?? I dunno, but it's true that faster way of typing is moving the wrist which is also more comfortable.
           My point is not about one TH "digraph", rather it's about T-H and R-L "positions" in my layout which has to be a high frequent combination contrary to the belief  many have established. These aren't same finger positions in short...

SpeedMorph said:

The Z underneath E seems like a waste of a position that, while not especially easy to type, still deserves something more common. I would probably want to switch the K and the Z, except that the KE/EK digraphs will increase same finger.

You are right, that position requires something more common, and again switching K and Z would increase same finger. So now this depends on which contributes to  speed more, optimizing for the digraph frequency or letter frequency. It has to be both of these. But I have a reason to place it this way. That is, placing an uncommon letter in a relatively easier position is not wrong, what matters is if we are placing a more used letter under a difficult spot. In that respect I've an idea of my own which i'll list later. If we are placing all letters according to their frequency and ease of hitting, I feel we would get only a few layouts to talk about. Rather we have almost infinite layouts 'cos we give some tolerance level. So here, we have to think about which are the characters that really deserve the Z spot in my layout; and then see if those characters are given any worse place. That's the way to go. B is one such letter I've misplaced in my view which is not a serious glitch in my opinion. Moreover the difference E-K would make from E-Z is huge. So i would stick with digraph theory..


My take on some popular (among us :P) layouts with regard to positioning :
Maltron :
serious glitch - L

Dvorak:
serious glitch - I, L
minor glitch - F, R, Y

Arensito:
serious glitch - W, M, Y
minor glitch - C, P

Colemak:
minor glitch - C, .

Klauser:
serious glitch - C
minor glitch - W, P, .

Capewll:
serious glitch - H
minor glitch - C, F, .

MTGAP2:
serious glitch - Y
minor glitch - .

[your 3.5 version is too good in this regard imho, but some I could find Y on left pinky on INESRTAO layout on your site! ]

mine:
minor glitch - B


So this is how i gauge. Hope i could drive the point.



SpeedMorph said:

This is a fairly minor point. Most keyboards are almost exactly as good on either. The biggest difference is probably the QWERTY 'b' key, which is a lot easier to reach on a good ergonomic keyboard than on standard. I designed MTGAP 2.0 (and most of the other versions) for standard keyboard but they are nearly as good on an ergonomic keyboard.

Yup, still there's a difference in design principles. And that gives room for a debate when u get slightly different scores. I tell you, the difference these minute dissection of typing characteristics yields you is silly. after all we know Dvorak, or Colemak or MTGAP or any modern layout is as good as the other; all much better than qwerty. It's the minute details we are indeed talking about which anyway is a fair point still. Same applies to how effective the same layout is on both conventional and an ergonomic one. Of course it's the same more or less.
     On a conventional keyboard, stretching would be effortless, but quicker would be bending(bottom row typing) for index finger as its a short finger. Similarly more error free would be typing on bottom row with pinky (u got to type very differently when pressing with pinky on top row as its a short finger. Touch typing is not possible without looking the keyboard and moving your whole hand and hence all the hassle. Error prone in my opinion)  though its easier and comfortable to stretch with it than bend. You don't face these hassles on a curved design where easier and quicker would be to type on top row. The weightages in algorithm changes hence in both designs. I would copy paste some differences I posted here before..
•    I left the spots under H and L blank as in my ergonomic design, there's no room for those spots which will collide with the thumbs (mind you, H and L doesn't fall in straight line, instead they curve downward in a suitable stroking fashion along the home row; so the spots under H and L fall a bit more downwards which will collide with thumb. It's a subjective thing, so leave it to my personal preference).
•    I try to keep hurdling the least possible (of course I can design a layout with even less home row jump, but loses out on other facets then). The reason is, a home row jump is twice a hurdle compared to same finger usage. It is not true on a standard keyboard as we already incline our hands on horizontally aligned letters, so stroking with index finger on bottom row and with ring finger on top row in succession is easy. But in my already curved ergonomic layout, a same finger is just half the hurdle as that of a home row jump (which moves twice).



Thanks for appreciation friend, truly heartening. You are after all a super star among few others having developed something ;) I could talk to others as well :)

Here's my explanation:-
1. I dunno how to reach the best performing layout unless we use a large cluster to "super-compute" the whole layouts existing :)Some big guns can/may  help :O.  'cos I feel these random evaluation may not yield the best always. There are chances they escape our attention (I understood what simulated annealing is after reading you people, never understood during my graduation :P). Yeah but, the layouts possible need not be 30! or 26! of course. We can reduce it greatly by applying constraints. We can talk more if needed.

2. Algorithm - haha, i feel its not really right to give some random numbers as cost or penalty to settle on diagraphs n positions. That can always vary from person to person. Cant we really calculate the time required in stroking keys from a pool of decent strength, normalize and then apply to algorithm accordingly? I'm pissed off when i see a cost of 50 for same finger and only 10 for home row jump :) This is how i differ. A and D on qwerty isn't inward roll IMO, so goes the differences. the way to calculate it is only by having real world data and employ a "just" penalty methodology. I didn't mean I know these values. It's just my idea to tackle which makes it incontestable, what say?? I know all these takes some patience and hard work which wont happen until we devote our whole time to it, but who has.. Not something we can finish off as a part time hobby.

3. And the corpus, whew.. This is something that caught my attention from long back and i dunno if there's been a study made like this, ie, take samples from real world excluding the non-english terms. I feel non-english terms disturbs the frequency pattern greatly and this is where differences arise. Of course a good size has to be ensured. The more the better. 100MB would do just perfect perhaps. I think the idea is conveyed. And of course its just not about that, the quality and nature of text matters. Say for instance, programmers don't type that fast. It's the reporters and likes who do. Incorporating programs in corpus disturbs the values with terms like autoform, sqrt,nodename what not.. which are not english terms. Fortunately they dont code so fast for a long duration. They do,if ever, in short intervals which they can do on any keyboard. I mean we dont have to optimize the whole system for them if that alters the pattern. The punctuation and symbol frequency is what that should be calculated from programs. So and so. If you could leave a word or two, we can talk more :)

Last edited by nawfal (06-Dec-2010 23:20:17)
Offline
  • 0
  • Reputation: 0
  • Registered: 03-Jul-2009
  • Posts: 189
nawfal said:
SpeedMorph said:

The TH digraph is a unique solution and it has its advantages but I don't think it's very elegant. Touch typing works a certain way; why should you deviate from that just for this one digraph? I'd rather use a keyboard that keeps things simple..

I dunno how dis/non-elegant TH combination makes in my layout. I would love to hear what you have to say on this. There's hardly anything that disturbs the regular way of touch typing, except the fact that touch typing is made more efficient.

Hi Nawfal,

I think only prolonged usage would tell.  It takes a lot of time to really "feel" a layout.  I remember having some issues while learning Dvorak which do not matter anymore.  On the contrary, I now feel they were not glitches at all, but the right way.  Just like beginning Colemakkers experience discomfort about the placement of R an S.

nawfal said:
SpeedMorph said:

The Z underneath E seems like a waste of a position that, while not especially easy to type, still deserves something more common. I would probably want to switch the K and the Z, except that the KE/EK digraphs will increase same finger.

You are right, that position requires something more common, and again switching K and Z would increase same finger. So now this depends on which contributes to  speed more, optimizing for the digraph frequency or letter frequency.

Nawfal, you haven't published statistics about your layout, have you?

nawfal said:

Dvorak:
serious glitch - I, L

I wouldn't say "I" is a glitch in Dvorak.  Of course Dvorak could have swapped "I" and "U", but having got so much things right in his layout, I'm sure he had his good reasons to do so.  Disclosure: I have "I" and "U" swapped into my customized Dvorak, but that's because I have to type Italian too, in which vowel digraphs involving "I" are frequent.

EDIT:  Regarding same-finger statistics, I think that the weight of same finger depends on where the same-finger occurs.  For instance, broadly speaking, I think that going from home row to top row is less disrupting than going from home row to top row.  Going from "E" to "," (Colemak) is the least cumbersome of the same-finger digraphs, I think.

Last edited by spremino (06-Dec-2010 10:22:12)

Dvorak typist here.  Please take my comments with a grain of salt.

Offline
  • 0
  • Reputation: 0
  • Registered: 02-Nov-2010
  • Posts: 38
spremino said:

I think only prolonged usage would tell.  It takes a lot of time to really "feel" a layout.  I remember having some issues while learning Dvorak which do not matter anymore.  On the contrary, I now feel they were not glitches at all, but the right way.  Just like beginning Colemakkers experience discomfort about the placement of R an S.

You are right, the feel you get is important, but that's bound to happen more or less. The poignant question here would be " is there any modern layout which lost out just because there was the "feel" missing??"... As I said, all of these modern layouts are too good and the benefits their proponents claim are marginal. We have millions of qwerty users and they don't ever complain. If they don't, are we expecting users of Dvorak or Colemak or anything like that to complain about something seriously? But some do. And they are being so critical of the minutest details. That's what. Dvorak users get used to its L, and it's too less a problem compared to what qwerty offers us. Same applies with I. It's not the best thing to place one of the most used keys outside home position. Now that wouldn't matter much, but still it's not the best idea. This is what I'm trying to convey. Getting the theory part right is very important before waiting for practical results. Practical results wont be so substantial to our senses to be it anything palpable. It's hard to gauge that way. Still having L and I under wrong spots is bad right. This is why i believe Colemak is a better layout compared to Dvorak, however comfortable a Dvorak user is with it. The feel falls in place with time. If 99% of qwerty users don't complain, if a nice share of Dvorak users don't complain about the 10th most used letter placed on right pinky, I feel keying one more key(T) when already moving to press a key(H) is just gonna be smoother.
   Come on, the gonna be effect of rolling over TH is just too small compared to pressing L or I on Dvorak. Pressing any key outside home position is undesirable. Why not press an additional easy to hit key when going for this undesired movement there by utilizing our effort to its maximum.
     

spremino said:

Nawfal, you haven't published statistics about your layout, have you?

I didn't get you. What stats? Sorry.


spremino said:

Regarding same-finger statistics, I think that the weight of same finger depends on where the same-finger occurs.  For instance, broadly speaking, I think that going from home row to top row is less disrupting than going from home row to top row.  Going from "E" to "," (Colemak) is the least cumbersome of the same-finger digraphs, I think.

Did not get you. Some spelling mistake right..

Offline
  • 0
  • Reputation: 0
  • Registered: 03-Jul-2009
  • Posts: 189
nawfal said:

We have millions of qwerty users and they don't ever complain. If they don't, are we expecting users of Dvorak or Colemak or anything like that to complain about something seriously? But some do. And they are being so critical of the minutest details. That's what. Dvorak users get used to its L, and it's too less a problem compared to what qwerty offers us. Same applies with I. It's not the best thing to place one of the most used keys outside home position. Now that wouldn't matter much, but still it's not the best idea. This is what I'm trying to convey. Getting the theory part right is very important before waiting for practical results.

I agree with the last sentence: do you know the theory behind the Dvorak layout?  I'd bet not.  Just like all the people who criticize it.  Dvorak obviously made his homework, and if he placed the "I" there, he must had had his reasons.  Now, if he had swapped "U" and "I", we wouldn't be here talking, but he decided otherwise.  Dvorak is an hell of a good layout, no matter what half-baked keyboard layout evaluation softwares say.  I'm seeing advantages of Dvorak that no other layout offers, which are often not acknowledged by evaluation software:
- perfect balance between hands (albeit not on fingers, and right thumb is unused)
- left hand is mostly static, making pressing the space bar very comfortable (the "chords and melody" approach: https://www.youtube.com/watch?v=YPfevKbJRmQ or the video which follows it)
- again, left hand is mostly static, which helps reducing the strain because of the non-natural shape of left side on standard keyboards;
- placement of "',." on left top row is perfect: they always follow and never are followed by a letter on the left home row, which would make for an awkward movement;
- my past experience as a guitarist taught me that alternating hands makes for faster sequences compared to same-hand rolls, when rolls are bidirectional.
- alternating hands makes your hands have time to relax between keystrokes.
- a side effect of vowels being separated from consonants is that Dvorak works well for typing languages other than English (as it is my case)

nawfal said:

If 99% of qwerty users don't complain, if a nice share of Dvorak users don't complain about the 10th most used letter placed on right pinky, I feel keying one more key(T) when already moving to press a key(H) is just gonna be smoother.

You forgot that 99% of Qwerty user do not know that alternative layouts exists, o are not interested anyway.  Ehi, as a software developer, I hit a lot more keys than "L" with my pinkies, yet they do not complain.  If one would like to make a really different layout, please consider moving control characters as well, consider distancing home keys (both indexes) more, consider using other keys for letters.   Then we would see some real innovation.

I considered moving from Dvorak, but no other layout made it worthwhile.

nawfal said:

Come on, the gonna be effect of rolling over TH is just too small compared to pressing L or I on Dvorak. Pressing any key outside home position is undesirable. Why not press an additional easy to hit key when going for this undesired movement there by utilizing our effort to its maximum.

I think you have misunderstood my comment as criticism, whilst - on the contrary - I have stated already that I think your "TH" trick is an interesting twist.  However, only prolonged usage would tell whether it would be confusing for users (I don't think so, but still, you have to check).

nawfal said:
spremino said:

Nawfal, you haven't published statistics about your layout, have you?

I didn't get you. What stats? Sorry.

I mean: home-row usage percentages, same-finger percentages, row-jumping percentages and so on.

nawfal said:
spremino said:

Regarding same-finger statistics, I think that the weight of same finger depends on where the same-finger occurs.  For instance, broadly speaking, I think that going from home row to top row is less disrupting than going from home row to top row.  Going from "E" to "," (Colemak) is the least cumbersome of the same-finger digraphs, I think.

Did not get you. Some spelling mistake right..

I wanted to say that not all same-finger sequences are equally awkward to type.  Some same-finger sequences are easier than others, and that should be taken into account.

Last edited by spremino (07-Dec-2010 01:21:00)

Dvorak typist here.  Please take my comments with a grain of salt.

Offline
  • 0
  • Reputation: 0
  • Registered: 02-Nov-2010
  • Posts: 38
spremino said:

I agree with the last sentence: do you know the theory behind the Dvorak layout?  I'd bet not.  Just like all the people who criticize it.  Dvorak obviously made his homework, and if he placed the "I" there, he must had had his reasons.

Now what exactly is the reason? The theory behind Dvorak is as public as it can get. As long as we don't have anything conclusive and concrete, we are not obliged to fall under their view. Atleast an explanation is what we are begging for. The only reason i can see for L is to reduce same pinky usage. And for I is to have OU combination right. Disregarding I for OU is not at all the way to go. Take any corpus, I comes a lot more than OU as a combination. Dvorak must be intelligent and he would have seen something else. But i tend to believe when something eludes us in spite of hundreds racking their brains, we shouldn't be relying on a 1930's man's brilliance neglecting all our intellect. As simple as that. Dvorak was successful in explaining his design principles. And we bought it not because we found his layout so easy to type, but because he could so beautifully explain it why. This is where I is a mystery. Same goes with L.

spremino said:

Now, if he had swapped "U" and "I", we wouldn't be here talking, but he decided otherwise.

We wouldn't be. So closing such gaps is the way to get something better. And  we are on course of finding "that something" which is always better right.  I talked only because he had not swapped. May be Dvorak was designing for a typewriter and he must have not given a fixed position for our hands, of course it doesnt rest somewhere when typing on a typewriter. So he could have expected our hands to move along the entire horizontal row, which is possible if u can visualise the way we type on a typewriter. So for him the entire home row would have been home positions. Which wont hold true for today's touch typing. I'm just seeing some reason. To err is human.

spremino said:

Dvorak is an hell of a good layout, no matter what half-baked keyboard layout evaluation softwares say.

We wouldn't disagree. But I don't think it's not possible to better it. I and L are paramount in this direction. And this is what, you find it really easy to type on it. That's the whole thing I'm saying too. These are all near optimal layouts on which typing is a breeze. Same applies to Colemak or mine. I said the differences are marginal. A Dvorak user would find it easy to type on Dvorak, a colemak user would find it easy to type on Colemak. I said these aren't palpable when it comes to level of pure personal experience. Now to pick one to be better, theories should come supporting it. This is where modern layouts score.
    And half-baked? A modern software could do all the possible evaluations in a matter of minutes what Dvorak has done whole his life. It can calculate a number of parameters and optimize it for all of them. A software generated layout is one that is the winner from a million layouts, how many could hav Dvorak tested? I can in a matter of seconds find a much more comprehensive frequency data than what Dvorak could find. Which are the principles of Dvorak modern layouts ignore? We are just maximizing Dvorak's own design principles with modern technology. Come on we respect the past one's for their great endeavor and the inspiration they are to us, but doesnt mean we just have to embrace them whole heartedly.

spremino said:

I'm seeing advantages of Dvorak that no other layout offers, which are often not acknowledged by evaluation software:
- perfect balance between hands (albeit not on fingers, and right thumb is unused)

I don't find it perfect. It emphasizes more typing on right hand, which is ok by the way. Workman has a 50-50 distribution. And it does no good because finger load distribution is more important. Now here is the answer for you - Dvorak emphasizes on bettering the load distribution to fingers and see what he has managed to. Now do we really believe our eyes or rely blindly on one man's obscure principles?? The same should have happened to I and L. He deserves all praise for innovating something this cool, but his successors have given us better. That's something happening across globe, across fields.

spremino said:

- left hand is mostly static, making pressing the space bar very comfortable (the "chords and melody" approach: https://www.youtube.com/watch?v=YPfevKbJRmQ or the video which follows it)
- again, left hand is mostly static, which helps reducing the strain because of the non-natural shape of left side on standard keyboards;
- placement of "',." on left top row is perfect: they always follow and never are followed by a letter on the left home row, which would make for an awkward movement;

Yup, now thats true. But are these major causes towards better typing? these are minor in my books.

Let see the scenario:

Very important - home row jump - winner Colemak
                         same finger - winner Colemak
                         distance traveled - winner Colemak
                         load distribution - winner Colemak
                         awkward spots - winner Colemak
                         inward rolls - winner Dvorak

Minor factors -   same hand - winner Dvorak
                        more home row typing for left hand - winner Dvorak

If u don't disagree with me on these, we both have just one winner here.

spremino said:

- my past experience as a guitarist taught me that alternating hands makes for faster sequences compared to same-hand rolls, when rolls are bidirectional.

Not bi directional, only inward rolls are faster, and error free compared to alternate hand typing. outward rolls are always undesirable  and hence minimized. Same applies to all other awkward same hand typing.


spremino said:

- alternating hands makes your hands have time to relax between keystrokes.

which is not true. Both are in effect the same. Only two characters are pressed with one hand before alternation. This is the better way.

spremino said:

- a side effect of vowels being separated from consonants is that Dvorak works well for typing languages other than English (as it is my case)

Good observation. But then I was concentrating on English :). Moreover many layouts have only one vowel along with consonants. All the rest lie on one side together.


spremino said:

You forgot that 99% of Qwerty user do not know that alternative layouts exists, o are not interested anyway.

And that's my point too. Had it been too tough on them, they would have pondered about alternatives. They didn't but. They are happy that means. 99% dunno about alternatives; add to that, though 100 pc would have wondered why the keys are jumbled on their keyboard, 98% wouldn't have even thought of possibility of having alternatives or they wouldnt have felt the need to think of something new. This is how Spremino we go by if we think about how comfortable people are. As much as you are comfortable on Dvorak like on no other, some hundreds are with Colemak. Now to think about which's better, theories matter. This is where Colemak "really is" better..

spremino said:

Ehi, as a software developer, I hit a lot more keys than "L" with my pinkies, yet they do not complain.  If one would like to make a really different layout, please consider moving control characters as well, consider distancing home keys (both indexes) more, consider using other keys for letters.   Then we would see some real innovation.

If you hit more keys with your pinky, and yet you don't find it problematic, you can realise how worthless the real world test is to find something better. To overload pinky than your stronger fingers is bad and which ever layout  doesn't do that is better. Period! It won't make that difference to your experience in real world, so is the case with our million qwerty siblings :) It doesn't matter to them.



spremino said:

I think you have misunderstood my comment as criticism, whilst - on the contrary - I have stated already that I think your "TH" trick is an interesting twist.  However, only prolonged usage would tell whether it would be confusing for users (I don't think so, but still, you have to check).

No friend, thanks for the concern though. I haven't. I seconded you too saying that's right. I was just saying I'm super confident if L, I and bad loading of fingers of Dvorak can go unnoticed, TH trick is just gonna be even smoother for users :)


spremino said:

I mean: home-row usage percentages, same-finger percentages, row-jumping percentages and so on.

I've, but I don't say it openly 'cos I know it isn't anything perfect but just the varying numbers you see with different evaluators. Moreover another reason is, I've seen drastically different specifications when different evaluators evaluate, though when we normalize we get almost the same thing. Do I need to really??


spremino said:

I wanted to say that not all same-finger sequences are equally awkward to type.  Some same-finger sequences are easier than others, and that should be taken into account.

Yup, and I've not considered that factor as it could become too complex to calculate different layouts. I did manually right, not with the help of software. It's important to see these factors. There's not any penalty system I've incorporated. Which's why I believe mine is not perfect layout :)

Offline
  • 0
  • Reputation: 0
  • Registered: 03-Jul-2009
  • Posts: 189

Nawfal, I was writing a rather long rebuttal point by point, but I gave up.

Talk is cheap.  Do you think you could beat Dvorak at his own game?  Then do it.

I myself think that you and other developers of keyboard layout evaluation programs are grossly underestimating the power of the human brain when compared to a computer.  Yes, computers can perform millions of calculations during the blink of an eye, but they cannot be smarter than what you teach them to be.  I'm not that strong on maths, but I'm sure a mathematical model of typing could be developed which would beat massive calculations.  Keyboard layout design is not rocket science, once you have your statistics about letters frequencies laid down and your goals clear.  If you had read the source code of Maxwell's keyboard compare applet, then you would have found out that he employed the most dumb algorithm ever to evaluate efforts (this is not to criticize his approach, for it works since it is performed by a computer).  I doubt Dvorak followed such a naive approach, because it could have taken him centuries to get the job done.  Since he got so many things right, things which modern layout designers fail to achieve, I trust Dvorak more.  I have yet to see these better than Dvorak's layouts (by "better" I mean a layot which achieves Dvorak's goals with less glitches).  I will welcome them, when they will come.

Have a nice journey into keyboard layout design ;-)

Dvorak typist here.  Please take my comments with a grain of salt.

Offline
  • 0
  • Reputation: 7
  • Registered: 21-Apr-2010
  • Posts: 818

Hope you don't mind me just chipping in here.

I really want to praise Dvorak.  But as I've mentioned on this forum, I still have pain in my neck and hands, and I can't rule out my layout.  Also I have only good experience in one layout.

I have been in an environment where I helped beginners get acquainted to computers and the Internet.  Those with and without typewriter experience are puzzled by the Qwerty layout (UK Qwerty in this case).  Punctuation is a real gotcha.  Not all people learn to touch type immediately, they can't be bothered - I put it off for over a decade!

Anyway the point I'm trying to get across is that they get loosely acquainted with Qwerty probably on a subconscious level.  You don't hear people arguing for a better hunt and peck arrangement apart from those new to the keyboard - who'd rather an alphabetical layout (which would have not been possible on mechanical writers).  Having said that I think Dvorak would be advantageous to the hunt and pecker.  I wonder if it would even harm Qwerty touch typists to have the layout printed in Dvorak as a western standard.

I've practiced a lot this year with my touch typing - probably too much.  Dvorak has become more comfortable.  I'm trying to spot troublesome words in Dvorak - but to be honest there aren't many -  I can only put that down to good design.  I have more issue with the weirdness of hand placement than anything else.

Taking the most frequently used English words (the top 500) there are only a few that I find awkward.  The word 'people' is one of them.  It's left hand heavy and involves an outward role.  I can already see 'start' being an issue in Colemak.

Regarding the pinky, I feel that I'm not doing Dvorak any favours with my alias.  I actually enjoy typing with my pinky these days.  There is a nice rocking motion for the right hand with words like: should and would.  These are even satisfying to type.  Having said that some people have pretty pathetic pinkies - I've seen some quite stunted little fingers - and these may prove difficult to type with.

I like the 'melody' approach of Dvorak - I might consider adopting the left thumb for spacing.  The left hand side of my body doesn't experience pain, unless I twist the wrist left a little.  I'm glad the hand doesn't do much.  I have no problem with the y or i.

I think my issue is more with my keyboard and typing style rather than my layout, but then again you never know.  My spelling lets me down probably more than the layout.  It's funny how I still repeat the same mistakes of old.

I personally think there is a better/simpler/faster alternative to the current crop of keyboards and layouts.  Input software could assist with word composition - and this could be used across your operating system.

I wouldn't be surprised if you could replace the left hand side of Dvorak with one button, and use software to guess and fill in the words.

There could be other tricks, like utilising the number row.  Or having keys that have duplicitous meaning.

Not wanting to dishearten those pursuing perfection.

(I am sure I would find the 'th' trick in nawfal's layout difficult at first, namely due to the gaps on my keyboard.)

Last edited by pinkyache (07-Dec-2010 19:33:03)

--
Physicians deafen our ears with the Honorificabilitudinitatibus of their heavenly Panacaea, their sovereign Guiacum.

Offline
  • 0
  • Reputation: 0
  • Registered: 03-Jul-2009
  • Posts: 189
pinkyache said:

I like the 'melody' approach of Dvorak - I might consider adopting the left thumb for spacing.

I recommend it.  It helps balancing more the load between your hands, too.

pinkyache said:

I think my issue is more with my keyboard and typing style rather than my layout, but then again you never know.

Could you post a video of your touch typing?

pinkyache said:

My spelling lets me down probably more than the layout.  It's funny how I still repeat the same mistakes of old.

Auto-correction software exists to help.

pinkyache said:

I personally think there is a better/simpler/faster alternative to the current crop of keyboards and layouts.  Input software could assist with word composition - and this could be used across your operating system.

I wouldn't be surprised if you could replace the left hand side of Dvorak with one button, and use software to guess and fill in the words.

It's being done already.  I don't remember how those keyboards are called.

Dvorak typist here.  Please take my comments with a grain of salt.

Offline
  • 0
  • Reputation: 0
  • Registered: 02-Nov-2010
  • Posts: 38
spremino said:

Nawfal, I was writing a rather long rebuttal point by point, but I gave up.

Oh, I missed it then. 'Cos I'm always critical of myself and would have loved to see what you have to say. More so when it's anything regarding keyboards, we only acquire valuable info which is valuable for all those who ponder over keyboards :). Moreover, as usual, I wonder what's wrong with what I said :) But to tell you friend, there's not one thing i say here with an intention to offend or be satirical. It's all healthy and constructive and I read posts here with a very neutral mind..

spremino said:

Talk is cheap.  Do you think you could beat Dvorak at his own game?  Then do it.

Talk in fact has too much potential, often gone unrealised. Yeah an ego-less open heart is paramount. I think it's always possible to beat/better. Do u think Shai can't beat Dvorak? or Michael? It's pointless to ask. There's always room; more so when it's no rocket science at all as you acknowledge. Bear in mind Dvorak was developed for a typewriter which is still hunt and peck typing in a way.

spremino said:

I myself think that you and other developers of keyboard layout evaluation programs are grossly underestimating the power of the human brain when compared to a computer.  Yes, computers can perform millions of calculations during the blink of an eye, but they cannot be smarter than what you teach them to be.

Actually I feel you are under estimating human brain. Computers are not meant do something smartly. They are meant to do something quickly. Humans are smart, and its the same smartness we feed into a computers brain. Computers can't be smarter than what we teach them to be, but we can be smartest to the possible maximum and teach a computer that exactly. To get the desired result in such short time. This is what programs do. Is there any human element that we cant incorporate in a program, may I know??I think everything is programmable depending on the quality of the programmer and this is where i don't underestimate the smartness of humans.

spremino said:

I'm not that strong on maths, but I'm sure a mathematical model of typing could be developed which would beat massive calculations.

I dunno how.

spremino said:

If you had read the source code of Maxwell's keyboard compare applet, then you would have found out that he employed the most dumb algorithm ever to evaluate efforts (this is not to criticize his approach, for it works since it is performed by a computer).

I personally found it not right too. That can be bettered. I didn't find MTGAP system to be perfect too, though it was better than Capewell's. No idea on how Colemak was developed. But its still possible for "us humans" to be smarter and make computer do those.. The point u make has nothing to do with the dumbness of a computer evaluation methodology, but it's about how less smart the developers of those programs were.

spremino said:

I doubt Dvorak followed such a naive approach, because it could have taken him centuries to get the job done.

Yes, this is where we have room for improvement. You stated that not all same finger sequences are equally awkward to type. Such complications make it very difficult to score a layout by hand. How many different possibilities would have Dvorak been able to evaluate, calculating the complete effect?? Add to that, how perfect would have been dvorak's frequency data?


spremino said:

Since he got so many things right, things which modern layout designers fail to achieve, I trust Dvorak more.

Which are those? I asked you which are Dvorak's design principles modern designers tend to ignore? I see only two things. One is same hand usage. I excuse it as they are designed with a different approach (quick roll) which I feel is a better approach. Second one is less movement for left hand away from home row, which IMO isn't a serious factor. What about Colemak, MTGAP, mine etc which has lesser same finger, lesser home row jump, better finger load distribution , more home row typing etc which are factors Dvorak himself wanted to maximize.

To me, a layout that doesn't distribute load uniformly doesn't qualify to even talked about. 'Cos its easier to get better technical specs by skewing load distribution

Offline
  • 0
  • Reputation: 0
  • Registered: 02-Nov-2010
  • Posts: 38
pinkyache said:

I have more issue with the weirdness of hand placement than anything else.

.
.
.

I think my issue is more with my keyboard and typing style rather than my layout, but then again you never know.

I second. The quality and shape of layout is too pivotal. I advocate, for myself and all, keyboards of laptops but without the staggered rows. It's the shape of the keyboard that obstructs touch typing as I could observe. You can see something here. It need not be perfect. I prefer the larger and less spaced out keys of a laptop for the keys shown in the below pic

Truly_Ergonomic_Keyboard.jpg

http://www.crunchgear.com/2010/10/11/th … tural-est/



pinkyache said:

(I am sure I would find the 'th' trick in nawfal's layout difficult at first, namely due to the gaps on my keyboard.)

Exactly. The shape of the keyboard is critical here. The spacing of keys is important and designing in the most ergonomic manner wont result in a uniform shape for all keys. If you drum on a desk in your natural style, you would see the ring and middle finger falling closer compared to middle and index finger. The middle and index finger are wider apart. So to pull off the TH trick without even the initial hiccup, I feel T and H should be a bit wide. You wont understand this if you try this on your regular keyboard 'cos from your subconscious mind fingers fall exactly on keys. Try getting finger marks when u drum naturally on a white paper :)

Last edited by nawfal (08-Dec-2010 08:36:50)
Offline
  • 0
  • Reputation: 0
  • Registered: 02-Nov-2010
  • Posts: 38

Q    G    U    F    .           ,    M   C    P    B
A    O     I    E    L          H   T    S    N    R
X    K     Y    Z                    D   W    V    J


To appease opposite hand typing brigade, I've created this. I don't think it's as good as my original. But I find this better than Dvorak. BR digraph would be a limitation but I've my point vis-a-vis Dvorak.

Few things:
1. Same finger is lower than Dvorak
2. Home row jump is lower than Dvorak
3. Distance traveled is lesser (more home position typing I mean)
4. Better load distribution
5. Better moving to center column
6. Frequent keys not under awkward spots.
7. More inward rolls
8. Worse outward rolls compared to Dvorak though.
9. Same hand typing (or hand alternation) isn't as good as Dvorak, but I can't help it further where I don't design clubbing alphabets along punctuations. Moreover i don't consider this a serious factor compared to other factors.
10. Another minor factor would be lesser typing for left hand away from home row. This layout fares almost on par with Dvorak as I calculate.



I would like to present my main layout this way.

V    Y    U    M    .            ,    F    C    P    Q
O    I     E    R    L           H   T    S    N    A
X    K    Z    W                     D   G    B     J

Makes more sense.

Last edited by nawfal (08-Dec-2010 09:41:54)
Offline
  • 0
  • Reputation: 0
  • Registered: 03-Jul-2009
  • Posts: 189
nawfal said:

Moreover, as usual, I wonder what's wrong with what I said :) But to tell you friend, there's not one thing i say here with an intention to offend or be satirical. It's all healthy and constructive and I read posts here with a very neutral mind..

Nawfal, nothing of what you said was wrong.  It's just that I realized keyboard layout design seems such a subjective science.

As you have learned, modern designers think same-hand rolls win over inward-rolls a-la Dvorak.  I think it heavily depends on the keyboard you are using.  On a laptop keyboard, whose keys are flat and have short travel, same-hand rolls work good.  But on a keyboard whose keys have longer travel, I have found that same-hand rolls are awkward.  Since shorter travel keys are deemed to be less ergonomic, I stick to longer travel ones.  In doubt, ask here: geekhack.org (select "keyboards" forum).

A similar issue happens regarding what the designer of Carpalx calls "stroke path".  That's what I thinking about when I talked about the "static left hand" in Dvorak.  Sequences like WA/AW, ES/SE, RD/DR (Qwerty), are difficult to type - less so on a laptop keyboard - yet both Colemak and your layout have them.  Dvorak hasn't, and that is a big win to me.  Carpalx fully optimized layouts have not stroke-path issues, but still I think are inferior to Dvorak regarding inward rolls.

All the above said, I don't mind Dvorak having a little more same-finger, awkward placement of L, or worse; it's still better than the alternatives.

nawfal said:
spremino said:

Talk is cheap.  Do you think you could beat Dvorak at his own game?  Then do it.

Talk in fact has too much potential, often gone unrealised. Yeah an ego-less open heart is paramount. I think it's always possible to beat/better. Do u think Shai can't beat Dvorak? or Michael? It's pointless to ask. There's always room; more so when it's no rocket science at all as you acknowledge. Bear in mind Dvorak was developed for a typewriter which is still hunt and peck typing in a way.

When I said "Talk is cheap", I wasn't dismissing constructive exchange.  What I was dismissing is this attitude of saying someone - Dvorak in this case - is wrong, without providing evidence.

I'll tell it with different words, maybe it will be clearer.  Do you think Dvorak could have done better had he relied on computers?  Show us this improved "Dvorak" layout.  To this day, I haven't seen a layout better than Dvorak (according to Dvorak's priorities, which I share).  How can one say his layout is better than Dvorak when Dvorak has 25% same-hand and his 35%?  Obviously the layouts have different goals.  How can one say his layout beats Dvorak when Dvorak's only difficult stroke-path is the least difficult RD/DR (Qwerty) and his layout has that plus all others?

nawfal said:

Computers are not meant do something smartly. They are meant to do something quickly. Humans are smart, and its the same smartness we feed into a computers brain.

Nawfal, maybe you forgot that Dvorak took years - ten if I remember correctly - to design his layout.  Now, I understand that computers can aid design, but I think it would be difficult anyway to beat Dvorak's dedication by working on your spare time.  In fact, nobody has managed to beat Dvorak to this time.  On the contrary, I remember reading about a guy who devised a genetic algorithm to select layouts according to Dvorak's design goals, and guess what: the algorithm ended up spitting out a layout strikingly similar to Dvorak.  So similar that the author acknowledged the difference could have been more from glitches into his typing model than from Dvorak actual shortcomings.  Dead link: http://www.visi.com/~pmk/evolved.html

nawfal said:
spremino said:

I doubt Dvorak followed such a naive approach, because it could have taken him centuries to get the job done.

Yes, this is where we have room for improvement. You stated that not all same finger sequences are equally awkward to type. Such complications make it very difficult to score a layout by hand. How many different possibilities would have Dvorak been able to evaluate, calculating the complete effect?? Add to that, how perfect would have been dvorak's frequency data?

Since Dvorak lacked a computer to throw layouts at, he had to be smarter.  I guess he started by carefully lying down language statistics (letter frequencies, digraphs and trigraphs frequencies) and after careful observation of accomplished typists, he started to laid down the letters.  It doesn't take a computer to tell you finger load: just add frequencies of involved letters.  I guess more advanced maths can help with more advanced issues.  Add to the recipe many years of development, and you have a dish which current layout designers have a difficult time to cook, in spite of having massive calculators at their disposal.

nawfal said:

I asked you which are Dvorak's design principles modern designers tend to ignore.

1 - Mostly inward rolls (except pinky to ring finger, where I find outward rolls to be better, at least on home row, and it seems Dvorak would have agreed with me);
2 - few difficult stroke paths;

1 is a moot point because modern designers think same-hand rolls are better.  2 however is a real issue.

I would like to see more creativity in placing keys.  Because of the difficulty of writing evaluation software, designers stick to Qwerty's separation of letters, numbers, symbols and control characters.  They stick to leave home keys - F and J (Qwerty) - where those are.  I'd like to see all of that challenged.  Moreover, I wish designer would share the way they designed their layouts, to help people who need a customized layout (as programmers do).

Dvorak typist here.  Please take my comments with a grain of salt.

Offline
  • 0
  • Reputation: 0
  • Registered: 03-Jul-2009
  • Posts: 189
nawfal said:
Q    G    U    F    .          ,    M    C    P    B
A    O    I    E    L          H    T    S    N    R
X    K    Y    Z                    D    W    V    J

To appease opposite hand typing brigade, I've created this. I don't think it's as good as my original. But I find this better than Dvorak. BR digraph would be a limitation but I've my point vis-a-vis Dvorak.

Thank you for sharing.  Hint: if you wrap your layout into a "code" tag - as I've done - it will displayed with monospaced font, making alignment of letters easier.  See: https://en.wikipedia.org/wiki/BBCode#BBCode_tags

Where are ':;- - especially ' -and other symbols used in regular texts?

Since you have compared this layout to Dvorak, I'm comparing it too: this layout has difficult stroke paths that Dvorak hasn't: GA/AG, OU/UO.  IF/FI matches Dvorak's EP/PE: which one is more frequent?

Could you please share some numbers about your evaluation of this layout?  It would help understand some choices.  For instance, I don't understand why you placed , in such a difficult position.  Maybe it is less frequent in your statistics?

Thanks.

Last edited by spremino (08-Dec-2010 12:20:46)

Dvorak typist here.  Please take my comments with a grain of salt.

Offline
  • 0
  • Reputation: 7
  • Registered: 21-Apr-2010
  • Posts: 818

I must admit I do fancy the look of that keyboard, but I'd want to split the thing in two.

The 'rainbow' approach certainly seemed preferable at first glance for me.

However I have a friend that plays the piano and he practices for about 7 hours a day.  It made me question why I was having issues with keyboarding when some piano players are quite energetic.  I'm not so sure that a static position is good.  My friends fingers and hand tremble occasionally but in the main he appears to cope.  I tried to find some piano playing tips on the web, and there was a suggestion to curve the fingers (more like a clenched fist).  If you do this (unless my touch typing has warped my hands), the fingers fall into line.  I used to type with a more flattened hand, but I'm beginning to appreciate the gains of having more of an arch (imagine clawing something).

Which brings in yet another variable - which would effect your choice of keyboard.

Your Dvorakish layout looks far more familiar to me!  Now I understand part of the rational for switching to a Qwerty variant like Colemak (for Qwerty touch typists.)

I will video myself at some time and throw it in another thread.

I know there are tools like auto correct.  Again operating systems really need to centralise this as a core service,  I want to train software and take those settings with me, i.e. dictionary.  We should be using our optimised editors for text entry even when we are on the web.  I digress.

Last edited by pinkyache (08-Dec-2010 14:28:25)

--
Physicians deafen our ears with the Honorificabilitudinitatibus of their heavenly Panacaea, their sovereign Guiacum.

Offline
  • 0
  • Reputation: 0
  • Registered: 03-Jul-2009
  • Posts: 189
pinkyache said:

However I have a friend that plays the piano and he practices for about 7 hours a day.  It made me question why I was having issues with keyboarding when some piano players are quite energetic.

Bad technique and bad keyboard, that's why. 

Regarding technique, as soon as you post a video of your typing, I'll comment on it.  Let me know when you do, since I do not hang on this forum often, and I only subscribe to topics to which I take part.  Again, look for videos of accomplished Dvorak typists (Barbara Blackburn, etc.), as well as the old video I posted before.

Regarding keyboards, an ergonomic keyboard would be ideal, but using one will break your ability to type on laptops.  If that's not an issue for you, good.  For me it is, thus I prefer a good standard keyboard.  If you are located in USA, by all means buy an Unicomp Spacesaver with buckling springs.  If outside USA, try to get your hands on an used clicky IBM Model M in good condition (it has buckling springs too).  These keyboards give you feedback while typing, just like piano keyboard do when you play.  That matters a lot.

pinkyache said:

I'm not so sure that a static position is good.

I don't understand what do you mean.  Of course you shouldn't force your fingers to be static, you should just avoid unnecessary movements.

pinkyache said:

I tried to find some piano playing tips on the web, and there was a suggestion to curve the fingers (more like a clenched fist).

Again, good technique matters.  You have just learned how good typing technique is like.

pinkyache said:

I know there are tools like auto correct.  Again operating systems really need to centralise this as a core service,  I want to train software and take those settings with me, i.e. dictionary.  We should be using our optimised editors for text entry even when we are on the web.  I digress.

Sadly, this is not the case.  Modern applications tend to rely on their own lousy editors to let you enter text, and you have to jump through hoops to overcome this.

For instance, here is a list of common misspelled words for the Emacs editor (which I use): https://www.emacswiki.org/emacs/autocor … bbrev_defs

Last edited by spremino (08-Dec-2010 15:03:32)

Dvorak typist here.  Please take my comments with a grain of salt.

Offline
  • 0
  • Reputation: 7
  • Registered: 21-Apr-2010
  • Posts: 818

@spremino,  thanks a again for your recommendations.

You have said before that good technique involves economy of movement.  What I meant by a static position, is that you lock up your body.  A piano virtuoso moves up and down a keyboard.  The gap between hands changes too.  I'd guess that this keeps you flexible - exercise is good!  There is much mention about piano technique - it has to be worked at, and it's more than a straight back at the computer.

Regarding typing heros: I found for the first time Ron Mingo the other day, and it's a joy to watch him type:

https://www.youtube.com/watch?v=xv9o6t1jnuc

--
Physicians deafen our ears with the Honorificabilitudinitatibus of their heavenly Panacaea, their sovereign Guiacum.

Offline
  • 0
  • Reputation: 0
  • Registered: 03-Jul-2009
  • Posts: 189
pinkyache said:

You have said before that good technique involves economy of movement.  What I meant by a static position, is that you lock up your body.  A piano virtuoso moves up and down a keyboard.  The gap between hands changes too.  I'd guess that this keeps you flexible - exercise is good!  There is much mention about piano technique - it has to be worked at, and it's more than a straight back at the computer.

Yes, I agree.  Indeed, an ergonomic chair which I'd like to buy is designed to keep your spine straight and your body in constant movement: http://www.ceccherini.com/images/ergonomia_balans.jpg

Ladies and gentlemen, let me introduce the Stokke Varier Variable Balans: http://www.ceccherini.com/shop_skeda.as … 01&L=I&cl=

It's an expensive piece of hardware, and I should stop being cheap and buy one.  I've tried a knock-off of this cheaper version: http://www.ceccherini.com/shopping-onli … sp?L=I&cl=

but I wouldn't recommend it: it doesn't make your body move.

Also consider using an hand exercise ball to make a bit of stretching during pauses: http://www.amazon.com/Cando-Gel-Hand-Ex … B000B69S5G

pinkyache said:

Regarding typing heros: I found for the first time Ron Mingo the other day, and it's a joy to watch him type:

https://www.youtube.com/watch?v=xv9o6t1jnuc

Great video, thanks!

Last edited by spremino (08-Dec-2010 16:38:31)

Dvorak typist here.  Please take my comments with a grain of salt.

Offline
  • 0
  • Reputation: 0
  • Registered: 02-Nov-2010
  • Posts: 38
spremino said:

It's just that I realized keyboard layout design seems such a subjective science.

I have to agree to that. But I feel there's a way to tackle this by learning typing pattern from a pool of users. Yup it takes time and dedication.

spremino said:

As you have learned, modern designers think same-hand rolls win over inward-rolls a-la Dvorak.

I did not understand at all. Badly want clarification :(. What's same hand rolls? bi-directional?


spremino said:

On a laptop keyboard, whose keys are flat and have short travel, same-hand rolls work good. But on a keyboard whose keys have longer travel, I have found that same-hand rolls are awkward.

Again, how is this? The travel distance is the same in both cases right? Your fingers fall on the same spots right. Or I did not get you?

spremino said:

A similar issue happens regarding what the designer of Carpalx calls "stroke path". That's what I thinking about when I talked about the "static left hand" in Dvorak. Sequences like WA/AW, ES/SE, RD/DR (Qwerty), are difficult to type - less so on a laptop keyboard - yet both Colemak and your layout have them. Dvorak hasn't, and that is a big win to me. Carpalx fully optimized layouts have not stroke-path issues, but still I think are inferior to Dvorak regarding inward rolls. All the above said, I don't mind Dvorak having a little more same-finger, awkward placement of L, or worse; it's still better than the alternatives.

This is gonna be elaborate. This is the heart and soul of our talk. And i want you to assimilate this perfectly. No more room for lengthy rebuttals and explanations. I'm sorry to say this that I'm fed up of making this same point over and over and again - major advantages of non-dvorak layouts - yet you point one minor advantage Dvorak layout possess.
            Now let me admit that a keyboard layout design is more than few spec's I have stated here. Acknowledging this fact, let me add that these( same finger, row jump etc) are some of the important aspects of typing that I have posted here. Of course it's the final score (or the fitness as more knowledgeable people say) that matters. Now this value is different for different evaluators based on what they feel is a better typing model. I personally feel there is just one best way, though. My point is ( as I stated in my 2nd post of this thread) I haven't designed this bearing all the awkward stroke paths (which would require a computer), but a few important ones, manually, and that a complete design might look radically different from what I could reach now which might even resemble Dvorak. But can Dvorak be better even after a full design, say for instance vis-a-vis a completed layout like Colemak? A big no. Sorry if I'm sounding harsh.

Here are few reasons why. In detail. I want a reply (if you do) precisely to the point I make and not just casual statements like "Dvorak has his reasons" "you think you can beat at his own game" etc. Otherwise this discussion is only gonna prolong. Here I go.

[tip:
1) characters I post here are located as in familiar qwerty positions.
2) when I say Colemak, I mean by it modern layouts like Colemak, MTGAP or the one I could roughly design]

1. A layout that doesn't distribute finger load uniformly in accordance with finger strength doesn't even deserve to be talked about. As a person who has devoted a bit of time into this area, I can tell you effort and speed doesn't go hand in hand "always"; and it's easier to achieve better technical specs if you can mis place a key (which won't be palpable for end user in real world). Dvorak is a layout designed for speed and effort isn't given its due. I might sound too bold to say that, but I believe it. Your middle finger is stronger than your ring finger which in turn is stronger than your pinky. A layout that doesn't maintain this distribution is not a good typing model. Period. Experiences from real world is worthless, as 98% qwerty users wouldn't have thought about alternatives. In other words, if Dvorak doesn't seem to be problematic for users, they would find it even easier on a layout that does a better job, although these minute percentage differences need not be palpable. Theory matters. Colemak is "really" better.

2. I said same finger is higher in Dvorak compared to Colemak. Now you can assess which is a more difficult stroke path - AQ/QA, SW/WS, ED/DE, HY/YH, UJ/JU, IK/KI or the one you mentioned - WA/AW, ES/SE, RD/DR ?? Former is same finger sequence on same hand, latter lot is different finger sequence on same hand. I feel both are equally difficult because of the awkward structure of our conventional keyboard and weird placement of keys. But on a smooth ergonomic design (which is what I've designed upon), latter lot is easier to stroke. Hence I would say a layout that minimizes same finger more is better than a layout which minimizes WA/AW, ES/SE, RD/DR. Hope you got it.

    Here's stats as I could calculate. Different evaluators give radically different values though normailzed figures would be more or less same. And my corpus is far from perfect, so I understand the underlying problems in these values. But difference with Dvorak is too much so that it still gives an idea :)
Same finger:  Colemak - 0.96%
                       Dvorak - 2.1 %
                   Workman - 2.1%
                     Arensito - 3.5%
                    MTGAP2  - 1.16%
                      mine     - 1.3%
           my Dvorak like - 1.54%

3. Home row jumping strokes are much more difficult than WA/AW, ES/SE, RD/DR on any keyboard. Compare ZQ/QZ, ZE/EZ, NY/YN, MO/OM to WA/AW, ES/SE, RD/DR. The BR/RB and BL/LB sequences in Dvorak adds to this. Your opinion? Colemak has a better percentage score than Dvorak.

                    Colemak - 0.5%
                       Dvorak - 0.78 %
                   Workman - 1.3%
                     Arensito - 0.98%
                     MTGAP2 - 0.63%
                          mine - 0.1%
           my Dvorak like - 0.12%


4. More home row typing is on Colemak. It means lesser finger travel distance over all. It means lesser effort.

5. Frequently hit keys shouldn't be in awkward spots (as in mine or Colemak). That is more effort. Letter frequency has a say on effort. Digraph frequency has a say on speed. An L on top row assigned for pinky means  every time you have to type a frequent letter, you have to move your pinky to hit it. That's more effort, though it won't be impeding speed. I would say Dvorak layout doesn't have the best mix of speed and effort. A better model is to first fix positions for characters based on their frequency and ease of stroking.

6. I see only my two layouts with better inward roll percentage than Dvorak though. Which is not a big factor anyway

7. What matters is outward rolls which has to be a minimum. Dvorak wins hands down.


[outward rolls are still not as important as row jump or same finger;
by inward rolls and outward rolls i mean rolls over adjacent keys]

8. Regarding your point on static typing, I feel total effort is what that matters rather than how quickly you can type with left hand. I must admit that there are not many layouts that stresses the importance of this.But I don't think this is a very serious point. Normally people are ambidextrous when typing. Anyway you can find that there's more home row typing for left hand in my two layouts compared to Dvorak, and they move only as much as Dvorak away from home row (on left).




spremino said:

When I said "Talk is cheap", I wasn't dismissing constructive exchange. What I was dismissing is this attitude of saying someone - Dvorak in this case - is wrong, without providing evidence.

So what's the evidence I can be providing? Bring in a Barbara Blackburn of Colemak?? I was right from start backing my claim with the theory involved in determining a better layout..

spremino said:

I'll tell it with different words, maybe it will be clearer. Do you think Dvorak could have done better had he relied on computers?

One hundred percent!! All the more when he would be relying on computers he would be designing one for computers as well :P [hint: "I" would be inside home position]. His frequency data could be much much better. He could have tried a lot more different combinations he believe can be more or less as powerful. etc. If he designed now, he would have faced the competition as well :P.. Kidding.. :)


spremino said:

Show us this improved "Dvorak" layout.

IMHO, Colemak, MTGAP. Hopefully I can contribute as well.

spremino said:

To this day, I haven't seen a layout better than Dvorak (according to Dvorak's priorities, which I share).

Though I can't be very sure, I would believe my Dvorak version is better. For the reasons i stated above. AW/WA, ES/SE is not as difficult as same finger (strictly on my ergonomic design where keys fall very comfortably under our natural position of drumming. Otherwise you are right here).
      Similarly home row jump is more difficult than AW/WA, ES/SE on any keyboard. These two are Dvorak's own priorities. I believe these two would have come on top for him too.

I tend to believe a layout which finishes off your work traveling lesser, with a uniform distribution of load, with lesser "more difficult" parameters is a better typing model, in short.

spremino said:

How can one say his layout beats Dvorak when Dvorak's only difficult stroke-path is the least difficult RD/DR (Qwerty) and his layout has that plus all others?

Hope I was clear when I said Dvorak has more serious stroke paths. Just reiterating.




spremino said:

Nawfal, maybe you forgot that Dvorak took years - ten if I remember correctly - to design his layout. Now, I understand that computers can aid design, but I think it would be difficult anyway to beat Dvorak's dedication by working on your spare time.

That's a fair point. I would agree to you saying that a manual design formed after yrs of hectic dedication can be better than something formed in spare time even if it involves computers. But I believe, Shai and Michael has done their research as well. 

spremino said:

In fact, nobody has managed to beat Dvorak to this time.

I have to agree Dvorak scores well in certain areas. And I stated before it's possible to get such results if we are not adamant on controlling the finger load distribution. I first formed a Dvorak like layout which performed better than what i posted here, but I canceled it right there when it had more work for ring finger compared to middle. In my books Dvorak has already lost before anyone could beat owing to this concern..


spremino said:

1 - Mostly inward rolls (except pinky to ring finger, where I find outward rolls to be better, at least on home row, and it seems Dvorak would have agreed with me);
2 - few difficult stroke paths;

You find outward rolls to be better? Oh my GOD. In fact ring to pinky o-rolls is the toughest of rolls I find to hit :)

Last edited by nawfal (10-Dec-2010 13:51:31)
Offline
  • 0
  • Reputation: 0
  • Registered: 02-Nov-2010
  • Posts: 38
spremino said:

Where are ':;- - especially ' -and other symbols used in regular texts?

I could place it on number row of qwerty. The shape I have in mind is a split shape, where number keys can be placed in between the two hands.

spremino said:

Since you have compared this layout to Dvorak, I'm comparing it too: this layout has difficult stroke paths that Dvorak hasn't: GA/AG, OU/UO.  IF/FI matches Dvorak's EP/PE: which one is more frequent?

I'm not talking too much again. Both are almost similarly frequent. IF/FI though is very slightly more.

spremino said:

Could you please share some numbers about your evaluation of this layout?  It would help understand some choices.  For instance, I don't understand why you placed , in such a difficult position.  Maybe it is less frequent in your statistics?

Thanks.

"," is less frequent in any statistics compared to Y and F. I would say F and Y are twice as frequent as ",". Dvorak has worse choice in that regard.. "," fits perfectly where I placed it, considering the frequency of "," and ease of hitting it at that position.
I hope i posted stat in previous post. Apart from that what I have is inward and outward roll percentage.

Offline
  • 0
  • Reputation: 0
  • Registered: 03-Jul-2009
  • Posts: 189
nawfal said:
spremino said:

As you have learned, modern designers think same-hand rolls win over inward-rolls a-la Dvorak.

I did not understand at all. Badly want clarification :(. What's same hand rolls? bi-directional?

https://colemak.com/Hand_alternation

I do not think Dvorak's main goal was hand-alternation: inward-rolls mattered to him more than everything.  If you had to perform bidirectional rolls, then hand-alternation kicks in better.  I tried Colemak when I was just learning Dvorak - so I don't think I was biased - and found bidirectional rolls awkward on my keyboard.  They worked on laptop keyboards, though.

nawfal said:
spremino said:

On a laptop keyboard, whose keys are flat and have short travel, same-hand rolls work good. But on a keyboard whose keys have longer travel, I have found that same-hand rolls are awkward.

Again, how is this? The travel distance is the same in both cases right? Your fingers fall on the same spots right. Or I did not get you?

Horizontal travel distance is not the issue here.  Vertical travel distance - e.g.: pressing keys - on a standard keyboard (staggered keys) is the issue.  Laptop-style keys - that is: scissor-switch keys - have shorter travel.  As key-travel shortens, so do the difficult stroke-path I mentioned.  Again, this is a subjective issue: DreymaR uses the same keyboard as I do, but it doesn't share my discomfort, it seems.

nawfal said:
spremino said:

A similar issue happens regarding what the designer of Carpalx calls "stroke path". That's what I thinking about when I talked about the "static left hand" in Dvorak. Sequences like WA/AW, ES/SE, RD/DR (Qwerty), are difficult to type - less so on a laptop keyboard - yet both Colemak and your layout have them. Dvorak hasn't, and that is a big win to me. Carpalx fully optimized layouts have not stroke-path issues, but still I think are inferior to Dvorak regarding inward rolls. All the above said, I don't mind Dvorak having a little more same-finger, awkward placement of L, or worse; it's still better than the alternatives.

This is gonna be elaborate. This is the heart and soul of our talk. And i want you to assimilate this perfectly. No more room for lengthy rebuttals and explanations. I'm sorry to say this that I'm fed up of making this same point over and over and again - major advantages of non-dvorak layouts - yet you point one minor advantage Dvorak layout possess.
            Now let me admit that a keyboard layout design is more than few spec's I have stated here. Acknowledging this fact, let me add that these( same finger, row jump etc) are some of the important aspects of typing that I have posted here. Of course it's the final score (or the fitness as more knowledgeable people say) that matters. Now this value is different for different evaluators based on what they feel is a better typing model. I personally feel there is just one best way, though. My point is ( as I stated in my 2nd post of this thread) I haven't designed this bearing all the awkward stroke paths (which would require a computer), but a few important ones, manually, and that a complete design might look radically different from what I could reach now which might even resemble Dvorak. But can Dvorak be better even after a full design, say for instance vis-a-vis a completed layout like Colemak? A big no. Sorry if I'm sounding harsh.

At last, you nailed it down, Nawfal.  Better typing model is a subjective matter.  "Subjective" does not mean such model cannot be agreed, it means you need a lot of feedback and observation to evaluate it.  Difficult stroke paths do matter to me, because I type on a buckling-spring keyboard, which I've found to be the better for ergonomic reasons.  To people who type on membrane-based o laptop-like keyboards, this issue will matter less.

nawfal said:

1. A layout that doesn't distribute finger load uniformly in accordance with finger strength doesn't even deserve to be talked about. As a person who has devoted a bit of time into this area, I can tell you effort and speed doesn't go hand in hand "always"; and it's easier to achieve better technical specs if you can mis place a key (which won't be palpable for end user in real world). Dvorak is a layout designed for speed and effort isn't given its due. I might sound too bold to say that, but I believe it. Your middle finger is stronger than your ring finger which in turn is stronger than your pinky. A layout that doesn't maintain this distribution is not a good typing model. Period. Experiences from real world is worthless, as 98% qwerty users wouldn't have thought about alternatives. In other words, if Dvorak doesn't seem to be problematic for users, they would find it even easier on a layout that does a better job, although these minute percentage differences need not be palpable. Theory matters. Colemak is "really" better.

We agree on this, but how does for instance Colemak score vs Dvorak?  Here we go:

Colemak finger load: 7% 7% 10% 18% - 19% 15% 9% 11%
Dvorak finger load: 8% 9% 12% 13% - 18% 13% 12% 12%

I don't see that much of a difference. See "All the Books Combined" at http://viralintrospection.wordpress.com … d-layouts/

Ehi, according to the same page, when it comes to pinkies-load Qwerty blows both Colemak and Dvorak out of the water:

Qwerty finger load: 7% 8% 17% 20% - 19% 8% 11% 5%

nawfal said:

2. I said same finger is higher in Dvorak compared to Colemak. Now you can assess which is a more difficult stroke path - AQ/QA, SW/WS, ED/DE, HY/YH, UJ/JU, IK/KI or the one you mentioned - WA/AW, ES/SE, RD/DR ?? Former is same finger sequence on same hand, latter lot is different finger sequence on same hand. I feel both are equally difficult because of the awkward structure of our conventional keyboard and weird placement of keys. But on a smooth ergonomic design (which is what I've designed upon), latter lot is easier to stroke. Hence I would say a layout that minimizes same finger more is better than a layout which minimizes WA/AW, ES/SE, RD/DR. Hope you got it.

If your layout targets ergonomic keyboards, you should advertise it as such.  Otherwise, you can't really compare it to others. 

Yes, on an ergonomic keyboard, difficult stroke-path issues are absent.


nawfal said:

Here's stats as I could calculate. Different evaluators give radically different values though normailzed figures would be more or less same. And my corpus is far from perfect, so I understand the underlying problems in these values. But difference with Dvorak is too much so that it still gives an idea :)
Same finger:  Colemak - 0.96%
                       Dvorak - 2.1 %
                   Workman - 2.1%
                     Arensito - 3.5%
                    MTGAP2  - 1.16%
                      mine     - 1.3%
           my Dvorak like - 1.54%

I do really think same-finger is over-rated.  It depends on which fingers you are considering.  Again, if you are targeting ergonomic keyboards, then metrics change.

nawfal said:

3. Home row jumping strokes are much more difficult than WA/AW, ES/SE, RD/DR on any keyboard. Compare ZQ/QZ, ZE/EZ, NY/YN, MO/OM to WA/AW, ES/SE, RD/DR. The BR/RB and BL/LB sequences in Dvorak adds to this. Your opinion? Colemak has a better percentage score than Dvorak.

                    Colemak - 0.5%
                       Dvorak - 0.78 %
                   Workman - 1.3%
                     Arensito - 0.98%
                     MTGAP2 - 0.63%
                          mine - 0.1%
           my Dvorak like - 0.12%

What do you consider "row-jumping"?  Same-finger row-jumping or different-fingers row-jumping?  In the former case, you have a point; in the latter case, again, it depends.

nawfal said:

4. More home row typing is on Colemak. It means lesser finger travel distance over all. It means lesser effort.

Home-row typing is overrated.  Other keys besides home-row keys are easier enough to hit.  The most easier are I, O, U (Qwerty), I think.

nawfal said:

5. Frequently hit keys shouldn't be in awkward spots (as in mine or Colemak). That is more effort. Letter frequency has a say on effort. Digraph frequency has a say on speed. An L on top row assigned for pinky means  every time you have to type a frequent letter, you have to move your pinky to hit it. That's more effort, though it won't be impeding speed. I would say Dvorak layout doesn't have the best mix of speed and effort. A better model is to first fix positions for characters based on their frequency and ease of stroking.

"you have to move your pinky to hit it": if you meant you have to stretch your pinky, that's wrong.  You shouldn't move your fingers, you should move your hands insted.  Anyway, typing distance is overrated.  Yes, less is more, but floating your hands some more time is less important than other metrics.

nawfal said:

6. I see only my two layouts with better inward roll percentage than Dvorak though. Which is not a big factor anyway

Could you please post measurements?

nawfal said:

8. Regarding your point on static typing, I feel total effort is what that matters rather than how quickly you can type with left hand. I must admit that there are not many layouts that stresses the importance of this.But I don't think this is a very serious point. Normally people are ambidextrous when typing. Anyway you can find that there's more home row typing for left hand in my two layouts compared to Dvorak, and they move only as much as Dvorak away from home row (on left).

Maybe there is a misunderstanding.  I don't think a static left hand is good, I think a more static left hand can be a result of trying to work around the shortcomings of standard (staggered) keys.

nawfal said:
spremino said:

When I said "Talk is cheap", I wasn't dismissing constructive exchange. What I was dismissing is this attitude of saying someone - Dvorak in this case - is wrong, without providing evidence.

So what's the evidence I can be providing? Bring in a Barbara Blackburn of Colemak?? I was right from start backing my claim with the theory involved in determining a better layout..

You should show us Dvorak 2010.  That is, a layout which achieves the same goals Dvorak had and achieves them better than Dvorak 1936.  You - and others - have said that Dvorak is suboptimal because Dvorak didn't have computers at his fingertips.  Thus, to prove your point you should devise such a layout.  Saying that your layout is better than Dvorak's, when you have different metrics, isn't match.

nawfal said:
spremino said:

I'll tell it with different words, maybe it will be clearer. Do you think Dvorak could have done better had he relied on computers?

One hundred percent!! All the more when he would be relying on computers he would be designing one for computers as well :P [hint: "I" would be inside home position]. His frequency data could be much much better. He could have tried a lot more different combinations he believe can be more or less as powerful. etc. If he designed now, he would have faced the competition as well :P.. Kidding.. :)

Again, have you any proofs of this?  Or are they just speculations?  I think they are.  Have you read "Typewriting behavior" by Dvorak?  I'd bet no.  You just read here and there that Dvorak valued hand-alternation and inward rolls, and that's it.  As I've shown you - and I've not even read such book - there is more to the Dvorak layout than that.

nawfal said:
spremino said:

Show us this improved "Dvorak" layout.

IMHO, Colemak, MTGAP. Hopefully I can contribute as well.

Nope: they are better only according to different metrics.

nawfal said:

I tend to believe a layout which finishes off your work traveling lesser, with a uniform distribution of load, with lesser "more difficult" parameters is a better typing model, in short.

I do too.  However, we disagree about what "load" and "more difficult parameters" are.

nawfal said:
spremino said:

How can one say his layout beats Dvorak when Dvorak's only difficult stroke-path is the least difficult RD/DR (Qwerty) and his layout has that plus all others?

Hope I was clear when I said Dvorak has more serious stroke paths. Just reiterating.

Examples and measurements, please?

nawfal said:
spremino said:

1 - Mostly inward rolls (except pinky to ring finger, where I find outward rolls to be better, at least on home row, and it seems Dvorak would have agreed with me);
2 - few difficult stroke paths;

You find outward rolls to be better? Oh my GOD. In fact ring to pinky o-rolls is the toughest of rolls I find to hit :)

See?  That's subjectivity.  I couldn't type AR or OI (Colemak) comfortably on a keyboard: fingers almost lock.  ART (Colemak) would be a killer ;-)  Which is more common, your feeling or mine?  We should poll users.

EDIT: I managed to fix the syntax errors which were forbidding to post the BBCode formatted text.

Last edited by spremino (10-Dec-2010 22:19:27)

Dvorak typist here.  Please take my comments with a grain of salt.

Offline
  • 0
  • Reputation: 7
  • Registered: 21-Apr-2010
  • Posts: 818

@nawfal, It seems to me that your parameters are changing.  I think therefore you need to revisit and clearly state your design goals.

Surely a keyboard design for something like a touch pad, flat or even a different physical keyed layout will make comparisons with 'typewriter' layouts difficult.

I guess you are designing a layout for use on some hybrid of the conventional keyboard.  With an idea that touch typists could move over to it?

(I personally would like to see the keyboard die in it's current iteration.  So you can send me a prototype ;) )

A design for a touch pad, using finger sliding may be infinitely more ergonomic than tapping keys (I wonder if it is?)

Could you utilise a computer to create a layout/input device that didn't evolve from the constraints of a typewriter (which was the basis of Dvorak's design)?

I think also there is a huge difference between typing and touch typing and care should be taken when using the term.

Regarding the pinky, I did read some anecdote on the web (yes I know that's not helpful) about finger strength and muscle.  Apparently there are no muscles in the fingers at all.  It's all tendons and actually for gripping the pinky is your strongest finger.  If that is true, you have to be careful when referring to finger strength.  I think you'd be better using the terms dexterity and fatigue.

Last edited by pinkyache (10-Dec-2010 15:18:10)

--
Physicians deafen our ears with the Honorificabilitudinitatibus of their heavenly Panacaea, their sovereign Guiacum.

Offline
  • 0
  • Reputation: 0
  • Registered: 05-Jan-2010
  • Posts: 91
pinkyache said:

Regarding the pinky, I did read some anecdote on the web (yes I know that's not helpful) about finger strength and muscle.  Apparently there are no muscles in the fingers at all.  It's all tendons and actually for gripping the pinky is your strongest finger.  If that is true, you have to be careful when referring to finger strength.  I think you'd be better using the terms dexterity and fatigue.

It might very well be, but besides gripping is not related to typing, the index and middle finger can get additional help from the thumb while gripping. Every pianist can confirm that the pinky is the weakest finger for pressing stuff down, though.

Offline
  • 0
  • Reputation: 0
  • Registered: 02-Nov-2010
  • Posts: 38
spremino said:

I do not think Dvorak's main goal was hand-alternation: inward-rolls mattered to him more than everything.

Hand alternation and inward roll technique are basically two techniques and giving either of them more importance is inevitable.
1. Dvorak designed for typewriters and it's understandable hand alternation is the desirable technique there.
2. Almost all the modern layouts designed on roll technique has vowels and consonants clubbed. And layouts with least same hand sequences have vowels and consonants on opposite sides. It's easy to conclude Dvorak employed latter technique. Though within the set of characters on one side, he tried to have more inward rolls than outward rolls which is good.

spremino said:

If you had to perform bidirectional rolls, then hand-alternation kicks in better.

I dunno if modern layouts are designed employing bi-directional roll principle. At least I haven't. Outward rolls are undesirable, but not as much as same finger or home row jump.


spremino said:

Horizontal travel distance is not the issue here. Vertical travel distance - e.g.: pressing keys - on a standard keyboard (staggered keys) is the issue. Laptop-style keys - that is: scissor-switch keys - have shorter travel. As key-travel shortens, so do the difficult stroke-path I mentioned. Again, this is a subjective issue: DreymaR uses the same keyboard as I do, but it doesn't share my discomfort, it seems.

I don't think laptop keyboards have different vertical travel distance. They are almost the same if not a difference of one or two millimeters. Rather the dimensions of keys are different where laptop keys are bigger with lesser spacing between keys which is better I feel. But both are basically not designed scientifically in accordance with where our fingers fall naturally.


spremino said:

Difficult stroke paths do matter to me, because I type on a buckling-spring keyboard, which I've found to be the better for ergonomic reasons. To people who type on membrane-based o laptop-like keyboards, this issue will matter less.

On any keyboard same finger typing and home row jumping( hurdling as Dvorak calls it) are more difficult stroke path where Dvorak comes inferior.


spremino said:

We agree on this, but how does for instance Colemak score vs Dvorak? Here we go:

Colemak finger load: 7% 7% 10% 18% - 19% 15% 9% 11%
Dvorak finger load: 8% 9% 12% 13% - 18% 13% 12% 12%

I don't see that much of a difference. See "All the Books Combined" at http://viralintrospection.wordpress.com … d-layouts/

You are right, Colemak doesn't handle this rightly. I've said this in my first or 2nd post of this thread. In that respect Colemak too doesn't qualify to be that perfect layout.

Apart from that, my statistics on this is slightly different. More or less same about Dvorak. But not with Colemak. One, right pinky isn't that loaded to be at 11%. Two, right ring finger isn't less used compared to right pinky.



spremino said:

If your layout targets ergonomic keyboards, you should advertise it as such. Otherwise, you can't really compare it to others.

I did it. But there are certain things universal. Like same finger ( On a standard design, I admit adjacent finger strokes are equally difficult which is what Dvorak was mainly concerned about), home row jump, distance, load distribution etc. These are most important in all which Dvorak is inferior compared to mine.


spremino said:

I do really think same-finger is over-rated. It depends on which fingers you are considering. Again, if you are targeting ergonomic keyboards, then metrics change.

This is how I could infer on it. Same finger is a bigger trouble compared to adjacent finger when it comes to stronger fingers like middle and index fingers. But on weaker fingers like pinky and ring finger, I feel same finger is easier to hit than AW/WA.


spremino said:

What do you consider "row-jumping"? Same-finger row-jumping or different-fingers row-jumping? In the former case, you have a point; in the latter case, again, it depends.

I've incorporated both in my above calculations. It's the whole total. Both are more difficult than AW/WA. As per my calculation, Dvorak has this almost 7 times compared to my layouts.


spremino said:

Home-row typing is overrated. Other keys besides home-row keys are easier enough to hit. The most easier are I, O, U (Qwerty), I think.

Can't help it friend if you find everything overrated than AW/WA. "I" is a very highly used key and if it's outside the home positions, it affects your speed. Besides I,O,U other easier spots are M,V,G,H,W,E,R (QWERTY)


spremino said:

"you have to move your pinky to hit it": if you meant you have to stretch your pinky, that's wrong. You shouldn't move your fingers, you should move your hands insted. Anyway, typing distance is overrated. Yes, less is more, but floating your hands some more time is less important than other metrics.

Either way, it's more effort. That's my point. This effort should be given to stronger fingers. It skews finger load distribution as well.

Here's my stat on inward rolls and outward rolls (quick rolling over adjacent keys):

Dvorak: inward - 5%
          outward - 1.14%

Colemak: inward - 4.4%
             outward - 2.235%

mine: inward - 9.7%
       outward - 3.1%

my dvorak like: inward - 6.4%
                     outward - 2.34%





spremino said:

You should show us Dvorak 2010. That is, a layout which achieves the same goals Dvorak had and achieves them better than Dvorak 1936.

People mostly nowadays believe goals different from Dvorak's are better for a keyboard meant for touch-feeding inputs to a computer - something nonexistent during Dvorak's time. It would be pointless to expect me - or others- to adopt typewriter goals.


spremino said:

You - and others - have said that Dvorak is suboptimal because Dvorak didn't have computers at his fingertips. Thus, to prove your point you should devise such a layout. Saying that your layout is better than Dvorak's, when you have different metrics, isn't match.

Now going by even Dvorak's game, I would just say the one i formed with lesser finger travel, more uniform distribution of load, lesser "more serious" stroke paths (home row jump) is almost better (just one - AW/WA, ES/SE - parameter alone wouldn't decide the winner). Well, I would just say it. Can't be too certain as I admit I've to get more authentic on frequency data, and varying stroke difficulties across fingers.

spremino said:
nawfal said:
spremino said:

I'll tell it with different words, maybe it will be clearer. Do you think Dvorak could have done better had he relied on computers?

One hundred percent!! All the more when he would be relying on computers he would be designing one for computers as well :P [hint: "I" would be inside home position]. His frequency data could be much much better. He could have tried a lot more different combinations he believe can be more or less as powerful. etc. If he designed now, he would have faced the competition as well :P.. Kidding.. :)

Again, have you any proofs of this? Or are they just speculations? I think they are. Have you read "Typewriting behavior" by Dvorak? I'd bet no. You just read here and there that Dvorak valued hand-alternation and inward rolls, and that's it. As I've shown you - and I've not even read such book - there is more to the Dvorak layout than that.

Now what's the proof mate I could be giving when the question itself is of hypothetical nature? I can only speculate based on rational and commonsense.
1. He would have had different ideas if he's learning from the video footages of how touch typing works. It's a bit different from typewriter typing where hands don't rest.
2. His frequency data would have been much more authentic.
3. And he could have tried millions of possibilities which would yield better layout on his own design principles. Hand calculation of 300 plus statistics on each layout itself would
be hectic so as to find the optimal one.





spremino said:

I do too. However, we disagree about what "load" and "more difficult parameters" are.

We both know what load is.. We agree on this, but how does for instance Colemak score vs Dvorak? Here we go: is something you yourself said. There's no doubt a layout that distributes evenly is better. And as far as more difficult parameters are concerned, I would say BL/LB, BR/RB and likes on Dvorak are more serious than AW/WA, ES/SE (qwerty). Same finger is almost the same as AW/WA and likes. And then there's the overall effort which is given by finger travel.

[

spremino said:

Examples and measurements, please?

The whole essay I was saying it. Refer to the specs I posted in the previous post.

spremino said:

See? That's subjectivity. I couldn't type AR or OI (Colemak) comfortably on a keyboard: fingers almost lock. ART (Colemak) would be a killer ;-) Which is more common, your feeling or mine? We should poll users.

Haha, I think it's possible to get real world data which would tend to converge to a point. I believe inward roll is easier comparatively. Something wrong with your fingers :P

Offline
  • 0
  • Reputation: 0
  • Registered: 02-Nov-2010
  • Posts: 38

@pinkyache, Design goals are simple. Give weightage for different stroke penalties and calculate the metrics for different layouts. A layout that tops is the best.. I would say stroking should follow an inward path and everything else with same hand is undesirable. So other than the inward rolls, combinations should be typed with opposite hands.
Here are my priorities in descending order:
1. Finger load has to be distributed with some equity. Load has to be perfect, then comes everything else.
2. Characters shouldn't placed in difficult to hit positions. We get better specifications if we can place a letter in an awkward spot and most layouts have employed that. It in the end that contributes to more effort.
3. Maximum typing has to happen on home positions. Which reduce the overall distance traveled.
3. Home row jump is what I feel the most impeding factor and hence I kept that the minimum possible.
4. Next comes same finger, to my fingers.
5. Then comes adjacent finger typing on different rows.
6. I preferred maximizing inward rolls then over minimizing outward rolls which I feel now was a tad foolish.


pinkyache said:

(I personally would like to see the keyboard die in it's current iteration.  So you can send me a prototype ;) )

I'll have to send u a rough sketch :) mail id? ;)

pinkyache said:

A design for a touch pad, using finger sliding may be infinitely more ergonomic than tapping keys (I wonder if it is?)

You are in a haste to move away from this whole conventional style of typing I guess :)


pinkyache said:

Regarding the pinky, I did read some anecdote on the web (yes I know that's not helpful) about finger strength and muscle.  Apparently there are no muscles in the fingers at all.  It's all tendons and actually for gripping the pinky is your strongest finger.  If that is true, you have to be careful when referring to finger strength.  I think you'd be better using the terms dexterity and fatigue.

I too read that. But how are we going to believe it when our pinky is really the weakest :P. It's not just dexterity, but the strength too that makes our pinky the worst choice for typing. I dunno what they say. And I have not read something very authoritative. Can you post link?

Offline
  • 0