Workman Layout for Vim | Hacker News Reader

Workman Layout for Vim

axiomatic.neophilus.net

21 pointsrcarmo68 comments

Last year during a couple month break I had between jobs I decided I would learn a new keyboard layout as it was something I always wanted to do, but never wanted to commit to the learning process while I still needed to be proficient at typing.

Your story is similar to mine. I also use Colemak-DH. However, I sort of quit Qwerty cold turkey. I also used to touch type about 80 wpm on Qwerty, and because I didn't maintain it or practice while learning Colemak, I completely overwrote my muscle memory. Overall, the layout is more comfortable for me, but hard to justify especially if you rather not tamper with default keybindings.

I'm also a vim keybinding user, but in Emacs evil-mode mostly. Vim keybindings are definitely made for qwerty, and to me not rebinding the keys just seemed insane.

I ended up spending a weekend, reviewed all the keybindings I use, and ones I should probably use more, then wrote it all out [0].

I remapped a lot of keys back to their qwerty positions, but I also took the opportunity to make some changes that I thought would be more ergonomic. I also came up with me own mnemonic system for the re-mappings.

For example:

  | function          | before | after | new mnemonic      | Commentary                                 | ------+-------+-------------------+--------------------------------------------| | g s   | search file at pt | need to free up `g f'                      | | g S   | search file.. etc | need to free up `g F'                      | | E      | F     | far WORD          | foot/forward are other possible mnemonics  | | g E    | g F   | far WORD rev      | foot/forward are other possible mnemonics  | | e      | f     | far word          |                                            | | g e    | g f   | far word rev      |                                            | | f      | s     | search            | right next to till :)                      | | F      | S     | rev search        |                                            | | v      | r     | range             | see note below                             | | V      | R     | range lines       |                                            | | C-v    | C-r   | range block       |                                            | | g v    | g r   | range restore     |                                            | | r      | v     | revise            | convert is another possible mnemonic       | | R      | V     | revise mode       |                                            | | `      | j     | jump              | easier to reach and now mnemonic           | | '      | J     | jump to line      | same key as j now, which makes sense to me | vim config [1] that I use if I find myself wanting to use (neo)vim. My evil-mode config [2] in Emacs. Remapping `less` keys [3].[0] https://github.com/willbush/system/tree/main/configs/keyboar...
[1] https://github.com/willbush/system/blob/main/configs/nvim/in...
[2] https://github.com/willbush/system/blob/82253534b92f3ab87d8e...
[3] https://github.com/willbush/system/blob/82253534b92f3ab87d8e...

I don't rebind my keys on Colemak for vim/evil. I got used to it pretty quickly. Beyond the basic hjkl movement commands I don't think anything is really meant for qwerty specifically. Remapping also breaks the mnemonics somewhat which I think is important. Emacs users got along just fine with Ctrl-n and p, so I see no specific reason remapping should be needed.

I learned colemak a long time ago, and progressed just like you, except I didn't try to retain qwerty speed. I figured I can look at the key legends anyways. I initially reached about 50-60 wpm but after that the speed still seems to progress slowly. Now after several years I can type about 100 wpm which I was never able to on qwerty, and I didn't do any significant practice on typing speed.

In my experience as a Colemak typist and heavy vim (or vim-emulation) user, the best solution is to bite the bullet and _not_ remap any of the keys. This is obviously annoying for hjkl navigation at first but you get used to it pretty quickly. Plus there's the added benefit of being able to easily use hjkl for navigation everywhere that uses them but _doesn't_ let you remap, which is a lot of places.

Yes. As a dvorak typist, hjkl being in eccentric locations hasn't bothered me.

In dvorak, 'hjkl' are still fairly convenient.'jk' are adjacent. 'h' is to the left of 'l'.

My solution has been putting important keys on a function layer so I can still reference them easily.

Totally agree.And I can use Vim without trouble on qwerty and on my own custom layout. The only adjustment to the layout I made when designing it was placing hjkl in reasonable positions (e.g. avoid the worst placements, but they're still not on the homerow).

How is it best when it's just bad? Why would you make the ergonomic of shortcuts worse when the big point of another layout is better ergonomics?

Because the economics aren't "just bad".The only shortcuts that might be worse on an alternative layout is hjkl, and you can argue that people way overuse them and it's better get into the habit of using other movement keys instead.
I've found that Vim ergonomics are way better on my own layout[0] than on qwerty, despite hjkl not being on the home-row.
[0]: https://www.jonashietala.se/series/t-34/

Yeah, "the only shortcuts" worse are the most used ones!And overuse argument isn't enough to even move them (it's an argument to maybe move word-based movements to the home row instead of them, but not to just move them out), let alone split them!
Your layout isn't relevant when talking about Colemak, it's trivial to make Vim way better even on qwerty since the defaults are very bad, so of course there are layouts where these letters magically appear in more ergonomic position, the argument is that it makes no sense to make it worse just to stick to the labels

I agree. I quickly started to associate the commands with the letters, rather than the key locations. hjkl you can get used to after a while.

This[1] links to http://www. workmanlayout.com/blog/ which is a 404. (space inserted to avoid auto link generation).Going to the main page of that site reveals some articles from the past few years, but also has online casino and crypto links disguised as posts. E.g.:
> is your premier source for crypto insights, covering the latest news, trends, and guides on blockchain, cryptocurrencies, and cybersecurity. Whether a beginner or a pro, you can find valuable information and tips to stay ahead in the Web 3.0 era.
Weird and shady.
I keep seeing all these people fancy keyboards and interesting looking layouts except that it takes them six to 10 attempts to type anything correctly. Don't really get the point, but having got curious about this "workman" layout and regretted trying to find out.
I was thinking of a pane layout manager ... Anything like that around?
[1]: https://axiomatic.neophilus.net/workman-layout-for-vim/

Your first link is broken

Cyber criminals use wordpress vulns to inject spam into old blogs. That doesn't make it less weird and shady, but might explain why it's weird and shady.

I touch typed qwerty for about 20 years using really bad technique(full left hand two fingers on right that roamed). When I got my first split keyboard I switched to Colemak and went through a bunch of touch typing tutorials to teach myself proper form. Without switching layouts I don't think I could have forced myself to start using proper form in qwerty.

Same as you. I switched to ColemakDH for a while and noticed that it improved my Qwerty typing speed.

It’s an old blog post from over a decade ago (2013). it originally pointed to: https://workmanlayout.org/

This is a problem with alternative layouts - you have to relearn or remap all the shortcuts for different apps as well as learn the new alphabet locations.Shortcuts sometimes use mnemonics, eg p to print or whatever, but they're often just in particular locations, like V for paste for example - it's just next to 'C' for cut - so if those two keys are separated on your new layout, it's confusing, and in any case, does it matter that it's 'C' for cut - just using whatever keys are in those two locations is so ingrained in muscle memory that they should be remapped.
The other issue is that most of these alternate layouts are optimized for writing prose, and sometimes make other things, code, for example, more difficult.

I use Colemak on my split keyboards (qwerty on my laptop and regular keyboards) and am a vim user. I just have a layer that has all my common programming symbols on my left hand and switches the homerow on the right hand back to qwerty so I can use vim navigation and it works seamlessly.

This is why something like colemak-dh is preferred. Keeps the "location shortcuts" where they are, but moves the others, and is by far the most comfortable writing experience I've had

I think ctrl-x, ctrl-c, and ctrl-v should have never existed. So difficult to press, for such common operations! While we still have prominent keys for rarely used functions, such as capslock, insert, etc.I have fond memories of those Sun keyboards with dedicated cut, copy and paste keys...

What's th best alternative you've seen?

Ehm... Sun keyboards with dedicated cut, copy and paste keys?

Hm, how is moving your whole hand off home easier for such frequent operations?

I very seldom write prose or long texts. 90% of my time is spent juggling emails, Excel tables or presentations. My right hand is on the mouse, the left one close to Shift, Ctrl, etc.

This is made for accounting people, they use to type on numpad with their right hand and ctrl -a -x -c -v with their left one for whole day.

> This is a problem with alternative layouts - you have to relearn or remap all the shortcuts for different apps as well as learn the new alphabet locations.Any alternative is going to have benefits and costs.
With learning alternative keyboard layout, "learn shortcut" mostly means you learn by the letter, not by the position on the keyboard.
It's least awkward for programs which you can drive with two hands on the keyboard (like vim), and most awkward for programs which want keyboard+mouse and have a bias for left-hand side of the qwerty layout for shortcuts.

> The other issue is that most of these alternate layouts are optimized for writing prose, and sometimes make other things, code, for example, more difficult.Why would it make writing code more difficult? Code is mostly just English plus punctuation, which most layouts don't change.

Code uses much more punctuation than prose.

I'm well aware, I did some analysis of my coding when I designed my layout[0].The point is that alternative layouts generally doesn't make it _worse_ than qwerty (maybe some weird esoteric ones might do, but the popular ones being discussed here are at worst equal to qwerty in that regard).
[0]: https://www.jonashietala.se/blog/2021/06/03/the-t-34-keyboar...

For the longest time I wanted to learn another keyboard layout, in the 80s and 90s it was Dvorak, but vi motion keys kept me from doing it.  I didn't like any of the solutions (remap, switch to qwerty for command mode, whatever)...  But I always felt like I was "leaving money on the table" by not learning something better.I've decided somewhat recently that I'm just fine with qwerty, any money that I'm leaving on the table is probably fairly small because I think slower than I type...
Some of that change in perspective may have been a coworker who used Dvorak, and watching that dude type was a seriously painful experience.

> because I think slower than I type...Do not confuse throughput with ping and not using visual cortex vs using visual cortex.

First time hearing about this keyboard layout. I wonder how much marginal efficiency there is to be gained.My hands' physical travels reduced by a lot ever since I learned vim. But I don't think optimizing fingers' travel would give as much boost in overall speed. On the other hand, though, you can argue that fingers travel between keys a lot more often than between keyboard and mouse.
What's everyone's experience? Was it worth it for you learning another layout?

Of course it worth it, but coleman/workman are for weak ppl who can not get rid of qwerty completely.

> What's everyone's experience? Was it worth it for you learning another layout?I designed my own layout to fix my RSI, which was very worth it. But I don't think it's worth it if you're after speed, the effort is better spent practicing typing IMO.
I wrote big lost about my experience here: https://www.jonashietala.se/blog/2023/11/02/i_designed_my_ow...

> as I remember Vim keybindings more on their associations (y = yank), than muscle memory of their original Qwerty locationsSeriously doubt this for such a frequently used key as yank, which is closer to hjk that require no word-based association at all

This makes no sense.It is _meaning_ that bounds certain Vim commands to certain keys, not the position of the keys on your keyboard. Moving the keys around doesn't change their meaning—"w" is still "w", for "word"—so making all those mappings is utterly pointless.
Unless one uses a layout different from that of their physical keyboard, in which case WTF?

I believe the issue boils down to qwerty-hjkl become workman-yneo.  hjkl don't have _meaning_, it's all about their specific location on the keyboard.  So if yneo is now left/down/up/right, you now need to do something else for yank/next/end/open.

^H, ^J, ^K, and ^L _had_ meaning and that meaning is why hjkl do what they do. Their location on the homerow of a number of layouts is an accident. And they are not very useful in practice anyway."Remapping" the whole keyboard around them is IMO a silly idea.

How do vertical tab and form feed relate to "up" and "right"?

You would have to ask to the engineers at Lear Siegler who made that decision many decades ago.That choice and the subsequent decision to print the arrows on the corresponding keys of a single model of their product line are why we got hjkl for cursor movement in Vi first… and now seemingly everywhere.

I would like to suppose this is what Home and End does on any modern keyboard.

Nope. .net/why-vim-uses-hjkl-as-arrow-keys" target="_blank" rel="nofollow noopener noreferrer">https://catonmat.net/why-vim-uses-hjkl-as-arrow-keys

"hjkl are arrow keys" because the arrows are printed on them and the arrows are printed on hjkl because pressing ^H, ^J, ^K, and ^L is what moved the cursor on the screen of the ADM-3A.

Yes, and H, ^J, ^K, and ^L is what moved the cursor on the screen of the ADM-3A because they are on the home row.

> ^H, ^J, ^K, and ^L _had_ meaningI can parse only High and Low which is what ^H and ^L do.

On the ADM-3A, moving the cursor on the screen was done by pressing ^H, ^J, ^K, ^L. The convention had been established by no-one-knows-who at Lear Siegler even before that specific terminal model came out.

I'm not the author but have been on Workman and Vim for about 8 years. I don't think the meaning matters at all outside of the initial learning phase. Now it's all completely "muscle memory"; having to think "I want to move a word along, word starts with 'w' and 'w' is the key for moving a word along" would annihilate any efficiency modal editing gives. When I'm using Vim, it's just a matter of "I want the cursor there and I know this particular finger movement does that".I remapped probably Vim 10-15 keys 8 years ago, no issues so far.

If anyone is interested on what I remapped, I deviated from HJKL for motion to JKLI (NEOI on Workman). I use space/shift+space for insert entry/exit, so losing I to a motion command was a nonissue. I then just mapped what NEO was to JKL. Wasn't a big deal surprisingly.

It's not all black or white. While most mappings are dictated by meaning, some are instead were chosen for their position, eg. hjkl, or relative position, eg. ? is the reverse of / just because in the US layout they are on the same key.

In reality there is very little meaning, there is just this false sense there is that keep all these bad defaults in place.Like, is it word forward or word backwards? W doesn't tell you.
So if you're used to qwerty W it makes perfect sense to keep the much stronger muscle memory intact when switching layouts and use non-W in the same spot

It does make sense because Vim by default is forward looking, so `w` moves forward a word and the secondary effect is to move backward with `W`.As another example `diw` reads "Delete Inner Word", or `dap` reads "Delete Around Paragraph".
It's way easier to just learn the new new keys and rely on the mnemonics IMO.

Still doesn't, the forward looking part is misapplied for movements where you have both directions in separate keys (this principle could work for ops similar to tT (forward by default, backward shifted)), just like you can't explain why B is moving backwards by Words, not characters or something else (is there another important context you need to know that "vim is word looking by default"?). Or why it's not Below to move downAnd in your D example your skipped the first part - why is it not delete forward/backward by character/word? Why is it not Dump for paste just like there is this weird Yank for copy? Why not Do to execute some command? Or maybe D should do something tied to its looks. After all X is delete based on visual mnemonic of crossing something out
It's even easier to learn relying on physical position mnemonics (like of your index J is go down by line, then go down by page is also index, but lower, so M). Or have adjacent fingers move left/right just like they move in cursors
(Btw, letter based mnemonics are also not useful for most of the world since they're tied to English)

Just because you could apply a different meaning to a character doesn't change the fact that some meaning, once learned, is more helpful than only relying on physical positions.Combining meaning with a physical position is even more superior than purely relying on one approach.
And to me, it was far easier to switch layout while relying on mnemonics than on key position, and this continues to be true even when I add new keybindings now when I switch between multiple layouts. Maybe other brains works differently though.

> Just because you could apply a different meaning to a character doesn't change the fact that some meaning, once learned, is more helpful than only relying on physical positions.It does precisely that - because "once learned" muscle memory / spatial orientation would beat your conscious efforts to map ambiguous semantic rules every time since they're more common/primitive/intuitive ("move finger up to move up")
> And to me, it was far easier to switch layout while relying on mnemonics than on key position
So wait, you've seriously tried home row cursor movements in different layouts, then seriously tried moving to keycap-based cursor, and found home row keys to be worse despite their objective ergonomic benefit that the alternative layouts tout as their benefit over qwertys?

I feel like you're almost being deliberately obtuse when you're ignoring the context that it's about learning the movements in the first place.This isn't productive in any way.

> you can't explain why B is moving backwards by Words, not characters or something else (is there another important context you need to know that "vim is word looking by default"?).Yes, there is some concept inside of this. Some commands are singular such as zz, some are paired (allmost all), some are double-paired, and only one set of commands is triple-paired, this is the list:
w W b B e E ge gE
Take care to find the logic by your own, please.

I took care, found the logic seriously flawed, explained some of the reasons above. Take care to address the substance, please

I don't see your attempts to understand the logic I have written. Why don't to RTFM what is the connection between the 8 commands I have provided?

But I do see how you fail to understand the failure of this logic, and me reading the manual again won't help you get it. So why don't you correct the assumption that you're the only one who reads manuals and try to think about the substance of the conversation?

Listen me apprentice. You gave us the question about why b command goes backwards while most of backwards commands are capitalized with an assumption that it has to contain some logic unknown by you. You have received the logic from me. What fucking else do you want from me?

No, my ignorant pseudo-master, you couldn't even understand my question before trying to shrowd your supposed answer in the mystery of a manual: it wasn't about capitalization, but object type.So I've received nothing from you, bBwW doesn't cover it. And I don't want anything from you, it's pretty clear you can add nothing useful to this conversation with such a combination of arrogance and ignorance

Well, your clarification did not help me to understand what you are looking for, try formulating your question once more. You just told me that your question is somewhere higher.

Single-letter commands have a clear meaning once you have learned enough of Vim. https://stackoverflow.com/a/7411851/546861> Like, is it word forward or word backwards? W doesn't tell you.
W doesn't have to tell you anything. W being the reverse of w is a grammar rule that is supposed to be learned and internalized, like the other grammar rules (count, etc.).
> So if you're used to qwerty W it makes perfect sense to keep the much stronger muscle memory intact when switching layouts and use non-W in the same spot
No. What would make perfect sense would be to retrain your muscle memory around your new layout.

The clear meaning of "h only makes sense on the keyboard used by vi's author" is that it's meaningless for everyone but the vi's author. And "once you've learned enough" is self-fulfilling for any setW is not the reverse of w in my comment once you read that it's about direction
And it makes no sense to go through the great pain of retraining when you can't explain the benefit

I’m into some pretty obscure keyboard layouts, you’ve probably never heard of them.

Reply on news.ycombinator.com