stuff

[vvp]

kbdfr, CeeSA:
I would need it then I would prefer a system level "macro expansion application" for such a huge amount of macros.
The application would be started with a special key combination and would allow me to select the right macro based on the provided (typed in) tag. It would show available options and provide tab completion (similar approach like in dmenu).

The second not so good option would be a modification of keyboard firmware so that it can start macros based on tags (macro would be started in a similar way how compose key expands to funny unicode characters but all in keyboard firmware). But this would not have the nice advantage of seeing all the options to which a part of tag (which was typed so far) can expand.

Mouse is rarely needed with a good window manager (e.g. i3). And I can move from keboard to mouse and back without looking. So it is not the same. Or maybe you can find the top tipro keys without looking too. Then it would be the same
Anyway I do not want to talk you out of using the keyboard you like. Just my opinion about what I would like more.

[davkol]

derp

[CeeSA]

In my workingspace I have no rights to install anything...
Also I have 3 workstations but only one keyboard (KVM Switch)to work with.

[vvp]

You've just described Emacs. People call it an operating system for a reason.

Oooo no! I'm a vi user. Let's go fight which editor is better!

[davkol]

derp

[vvp]

You won ... provided that feature richness is the main criterion.

[jacobolus]

Big problem with emacs is that it was designed in a very ad-hoc way over decades, without nearly enough effort put into making simple, powerful abstractions. All the features are intertwined in ungodly and hard to figure out ways, and it takes a several-month-long hazing/initiation ritual to come up to speed with just the basics (this also makes emacs acolytes very devoted, just like painful hazing rituals in other organizations). Also, as a result, actually customizing things in emacs is orders of magnitude more effort than it should be, and all sorts of useful features aren’t worth bothering to build, because they’d be a huge pain in the ass to integrate with the rest of the giant spaghetti ball that’s already there.

On the flip side, the great thing about emacs is that other people (usually as a way of procrastinating from whatever real work they would otherwise have to do) have already put in a ridiculously huge amount of effort, so if you don’t really want to dive under the hood too much, there are a bunch of nifty ideas that have already been implemented, usually 20 years ago, and they work at least passably well. Even if someone else had to spend 2 months figuring out how to do something that would take an hour in a better system, as an end user, that work is already done and out of the way, so you don’t have to care.

And you can play tetris with it.

[vvp]

Well if the emacs discussion is serious then I doubt it can serve as a system level macro expansion utility based on macro names/tags. I think one needs to hook input event queue of every process for that.
E.g. I am in firefox preferences dialog and I'm changing the home page url and I happen to have that url under tag "<Macro>home" then I doubt that emacs can inject the URL corresponding to macro name "home" into the right edit box of the firefox process. I would be glad to be corrected if I'm wrong here.

[davkol]

derp

[kbdfr]

[…] the easiest is to have some kind of textual/typing based access: that is, give the command a name several letters long. Typing a whole word or even a few words is not only dramatically more efficient than moving the hands off the home row, it’s also easier to remember. […]

I just fail to understand why typing a whole command is supposed to be "dramatically more efficient" than pressing a single key, even if one has to leave the home row for that.

Suppose we want to highlight a sentence yellow in MS Word. We would first mark it (obviously using the keyboard for that).

• I would then stretch my left hand middle finger to hit the yellow key on the 7th row and then retract it while
• you would, if I understand you correctly, type something like "Ctrl-M yellow Enter".

I see no "dramatically" better efficience there, just a reasonably better one - for my approach.

I even more fail to see why typing "yellow" (or, for example, "insert pic caption") is supposed to be "easier to remember" than pressing a yellow key (or a key labelled "pic caption"). And if you shorten or abbreviate your commands ("yell", "piccapt"), they will of course be more prone to error or confusion and more difficult to remember, while a keypress remains a keypress.

But for anyone who doubts the power of limited keyboards to improve efficiency, just look to stenography keyboards, where with about 25 keys total it’s possible to type at several hundred words per minute. The logic is the same: make the motions small, with no reaching, and use combinations of fingers to encode an idea compactly.

Again, you argue with typing as a tremendously narrowed, almost olympic discipline: prolonged, text-only typing with the maximum possible speed (i.e. while someone is speaking). This has absolutely nothing to do with usual typing, where you are not compelled to keep pace with anything and where you will be pausing anyway to reflect things, turn pages, check your last action, turn the sound louder, take a sip from your coffee cup, have a look at another application or whatever in your specific work environment.
While of course optimizing one's workflow makes sense, there is just no need for your obsessive "making the motions small, with no reaching" as an exclusive rule in any situation.

[seebart]

davkol I will now give you a nice example of a (for today) strange layout that works well for me. This Commodore PC-10 keyboard is from 1985/86 and as you can see has zero modifiers next to spacebar. Control and Alt are nice and snug above left shift. Unused caps lock has moved way to the bottom right. The whole layout is XT'ish of course. So your initial question of "how many rows should a keyboard have? " is slightly relative as you can see even 30 year old keyboard layouts work for some people like me.

IMG_20150710_140456.jpg (923.73 KiB) Viewed 5044 times

[vvp]

Suppose we want to highlight a sentence yellow in MS Word. We would first mark it (obviously using the keyboard for that).

• I would then stretch my left hand middle finger to hit the yellow key on the 7th row and then retract it while
• you would, if I understand you correctly, type something like "Ctrl-M yellow Enter".

What jacobolus, me, and some others want to point out is that even weak touch typist can do about 50 wpm. That is about 4 characters per second. I can do that and I never ever tried to be a quick touch typist.
From that, the sequence <Ctrl>Myellow<Enter> takes 9/4 = 2.25 s. We want to say that we would not be able to look down at the keyboard to find the yellow button somewhere up there, move our hand from home row to the yellow button and press it and return our hand back to the home row. All that within 2.25 s. Because of that we think that limited keyboard is better than huge keyboard with a lot of buttons.
Of course there is some difference between my limited and jacobolus' limited. My limited keyboard is a bit bigger.

[davkol]

derp

[iandoug]

kbdfr: basically, you got distracted by 2 sentences of throwaway tongue-in-cheek teasing, and totally ignored the rest, the result of an hour of work where I was trying to answer your questions in good faith and at great length. Such a conversation is not worth it for me. Type on whatever you want.

Well I for one am grateful for the time you spent ...

I think the problem is that there are too many diverse use cases for a common instrument like a keyboard.

My needs as a HTML/PHP/etc programmer working in Kate in KDE are different to a programmer working in Vim or Emacs.
If we add in other users like gamers or secretaries or people writing screenplays or accountants or POS, then clearly one size is not going to fit all.

So for this debate to have any meaning, the original question needs to be modified to specficy the use case that the poster had in mind.

I found what you wrote interesting, as I am trying to design my own programmer's keyboard. I'm not wild about heavy chording, and prefer to have separate keys, with standard shortcuts still preserved.

[kbdfr]

[…] the sequence <Ctrl>Myellow<Enter> takes […] 2.25 s. We want to say that we would not be able to look down at the keyboard to find the yellow button somewhere up there, move our hand from home row to the yellow button and press it and return our hand back to the home row. All that within 2.25 s. […]

If, like you suggest, I needed 2.25 s for the whole sequence [go to key - press - return from key],
you bet I would have found another solution

Without being able to exactly tell how long it takes me to press even the most remote key on one of my lateral 64-key units and then return from it, I would say it is clearly less than half a second.
And the whole move doesn't interrupt my typing flow, but rather integrates it neatly due to the well-considered design in key clusters, the tactile orientation given by i) differently shaped keys and ii) the edges of the key fields,
and the fact that I have been touch typing longer than most of you can write at all

[vvp]

Amazing hand speed and precision.
Especially when you are not one of the younger ones.
In my opinion, you earned the right to use and preach Tipro.

[Muirium]

Come on, half a second is about right, so long as you know where a key is. Anyone taking 2 seconds out to hit F12 or whatever seriously needs to think about retirement! 2 seconds is an eternity.

For instance: it takes me 4 seconds to say that paragraph above. (I just timed it!) So 2 seconds is long enough for a whole lot of speech! It's longer than you think.

[vvp]

Come on, half a second is about right, so long as you know where a key is.

That is the important part. Especially when we are talking about 3x16 1u keys above the number row and without looking. I would say that means pretty good muscle memory. I know I would need to train for a long time to achieve some consistency with hitting the right keys. Not even talking about 0.5 second times.
Maybe I'm just clumsy compared to the rest of the of deskthority users

[Muirium]

I think you simply underestimate how long seconds are. 2 seconds is the kind of timing I'd expect from a 70 year old hunt and pecker. So long as there's some sense to the keys, you shouldn't have to read all the legends just to find any of them.

And I'm a 60% layout guy! For me, the half second journey is more than enough to annoy me. Half a second in motion, but more than that in disruption (engage Kbdfr quizzical eyebrow) in flow. Reaching for the mouse is a similar break. I take neither lightly! Though I'm sure he reckons everyone with a TKL or less is fumbling with the rodent every few seconds because we're too dimwitted to know how to move a cursor by keys!

[kbdfr]

Come on, half a second is about right, so long as you know where a key is.

That is the important part. Especially when we are talking about 3x16 1u keys above the number row and without looking. […]

Actually my keyboard is a bit bigger

Spoiler:

I admittedly often have to look (or rather glimpse), particularly when using the lateral units.
This is not in order to find a wanted key, though, as I know where they are, but rather to optimize my movement.

Which reminds me that it seems just silly to insist on moving arms, wrists and hands as little as possible
without completely banning any pointing device or at least using one which can be used without leaving the holy home row.

[vvp]

Muirium, of course you are annoyed when you need to hunt 0.5 seconds for a key. You are a 60% user. You hardly need to move your wrist to hit any key. And when you do not need to move your wrist then it is also very easy to be precise hitting any key.
But kbdfr needs to move his hand by about 6 cm when he wants to hit the top row keys. Which are still 1u keys as the rest of them. (And I'm not even talking about the keyboard extensions he has at the sides.) I think that requires quite a bit of training to hit the keys precisely and return back ... all in half a second. It is not so much about the time as it is about the proper muscle memory to achieve usable precision without looking.
Is it possible to hit the keys on Tipro precisely? Sure, I believe that. Pianists can hit keys at big distances too. But they are also training quite a bit. Compare that to me. I can hit keys precisely in about 0.25s and I did not need to train almost at all. My wrists hardly need to move at all when hitting any key at my Katy keyboard. I actually can type pretty comfortably with my wrists resting at the palm rests simply because my wrists do not need to move when hitting any key. For me it is easy to be precise and quick enough. Not that I would care about speed. Speed of typing is not important for my job.

[jacobolus]

Let’s imagine that I have a keyboard with 50 keys on it. Now imagine I can make chords using one special-purpose left-hand thumb key, and otherwise have 20 keys which I can press with my left fingers, plus 25 keys on the right fingers & thumb.

To limit ourselves here, let’s assume that every chord I make uses either zero or one left finger keys, plus one right-hand key.

So that’s 21 * 25 = 525 shortcuts, all of which can be pressed in a single motion using 2–3 fingers, all near the home position and without requiring any hand contortions. Or in other words, about three times as many macro opportunities as are available in the non-letter sections of that triple tipro setup. And that’s just 2–3 key chords unlocked by a single specific thumb key!

If we need more shortcuts than that our 50-key keyboard has effectively unlimited chords available. Even better, we can very easily organize such macros into groups and place those on specific combinations of fingers, and if we design the scheme carefully it is pretty easy to remember them / easy to avoid mistakes.

I promise you I can press a sequence of e.g. left thumb key + left index finger key + right ring finger key and release them, and then get back to whatever else I was doing much, much faster than someone else can jab a 1u key in row 7. And I promise if we need to do the same action 20 times in the course of a few minutes of typing, under some kind of stress or speed pressure, my way will be more accurate to boot.

[jacobolus]

As for pointing devices: first, it is great to have at least one analog input device available without leaving the home row by more than a couple centimeters, whether that’s a trackpoint, touchpad, trackball, small joystick, knobs, scrollwheels, whatever. Even better if there are multiple analog inputs available.

But more generally, there are many tasks where it makes sense to use mostly analog input with one hand (either a mouse or e.g. drawing with a stylus on a tablet), and leave the digital input to the other hand. There’s nothing impossibly bad about having a dedicated analog input device separate from the keyboard. Even with 25 keys available to the other hand, it’s entirely possible to develop a reasonably efficient one-hand chording typing scheme (30–40 words per minute or better; the simplest method would be to use Matias’s half-QWERTY scheme or similar), and 25 keys on one hand is enough for thousands of potential keyboard shortcuts, though it might take slightly trickier hand motions than in the two-handed case.