So, to summarize, the original typewriter layout (1868) was alphabetical, with the top row (A-N) going left-to-right and the bottom row (O-Z) going right-to-left. Then (1870), the vowels (including Y) were pulled out and put in a separate top row. After that (by 1872), changes were made in order to better support the use case of people receiving Morse code, and that's when we finally start to see something that looks like QWERTY. Additional changes later got it to the modern form, but by 1872 something QWERTY-like was in place.

And yeah -- if you look at the bottom two rows of a QWERTY keyboard, you can still see what remains of that alphabetical ordering, being left-to-right on one row followed by right-to-left on the row below!

In the next century, researchers will discover that the GUI wasn't designed to make computing harder by forcing people to find cryptic little symbols, randomly arranged on the screen, and break routine operations into tiny sequences of manual steps. And it wasn't called a "personal computer" because it turned each person into a computer.

"A common misunderstanding is that GUIs were invented to give ancient computers enough time for processing by slowing down user input speed. However our research shows that counterintuitively input latency was better at the time when GUIs were invented and then gradually got worse..."

This reminds me of Motel of the Mysteries :^)

"The legend was referred by Prof. James V. Wertsch,[22, 23] a professor of the Department of Psychology, Clark University, then it was regarded as an established theory in the field of psychology. "

The reproducibility crisis struck early, it seems.

Ha. so the reason that I is next to 8 is that early typewriters used the I as a 1(no independent 1 key) and the morse transcription company wanted to type years(1871) quickly. I love it.

Whatever its intent, QWERTY definitely hasn't impeded the fastest typists, who can regularly exceed 200wpm these days.

Odd to see no mention of the Linotype layout, also known as the "Etaoin Shrdlu", given that was also a common competing keyboard layout in that era.

Humans do not have fins, but Micheal Phelps can still cut through water. That elites can thrive is not a compelling argument when most people just want technology to get out of the way.

An alternative layout with commonly used symbols on the home row makes the QWERTY deficiencies immediately apparent. Significantly less effort required for writing prose when using something like DVORAK.

I really don't think people who type slowly do so because of QWERTY. Anecdotally, my dad basically isn't able to develop muscle memory for the key locations and will frequently revert to the "scan and then press with one finger" method. You could give him any layout and he would still type slowly. Pretty sure even an alphabetical order would trip him up, because it'd need to be broken into multiple rows, so he'd need muscle memory again.

And while this is speculative, given how close typing speeds seem on a cursory search between layouts, this suggests to me that the vast majority of the performance comes not from the layout, but from touch typing and effective use of multiple hands and fingers at the same time. All layout agnostic skills.

This is not to say that on an input method level, things cannot be further improved. I sometimes see stenography [0] related software and demonstrations on YouTube for example. It also isn't to say that there cannot be a benefit health wise (i.e. ergonomically) to alternative layouts. It's just that for speed I'm not convinced it affects much, and so I think it's the wrong thing to try and change. Especially considering that sometimes things that are suboptimal can be better by being the standard.

[0] https://en.wikipedia.org/wiki/Stenotype

Dvorak is way more comfortable and is very fast for me, but its emphasis on alternating hands causes me to frequently invert pairs of letters when typing fast, probably because my left and right brain isn't perfectly coordinated.

I have considered switching to Colemak, which is supposed to have less of that, but as a 100+ wpm typer of Dvorak it really is diminishing returns.

I got into alternate keyboard layouts and developed my own (roughly an optimized NIRO). When I tried using it on my small Surface Go I found that my fingers would 'jam' typing letters close together, so I leave that in QWERTY so it happens much less.

Yeah, there is a conflict between "use your very best fingers for everything" versus "spread the work across fingers so that it gets done sooner."

I tend to favor the latter.

Everyone has their own style I would assume. I've been on a PC for 25ish years now and I still type with 3 or 4 fingers almost all the time. I've tried to be more disciplined but I manage 100ish corrected WPM and it gets the job done, I never find that I'm limited by my typing speed.

We've had a few conversations on HN about how QWERTY may not be optimal for mobile swiping, but it's better than the layouts optimized for typing. Dvorak, with all the vowels on the left hand home row, would be terrible for swiping, what with all the English words just a vowel away from each other.

As it is the UIO vowel complex is a frequent problem with swiping. I'd like to see U and O moved away from each other, in theory.

Yes, normal Dvorak is not great for use on a smartphone, unless you usually type with both thumbs. For one thumb, especially swiping, I've found one-handed Dvorak for right hand to be a great fit. Works well for me on Multiling O Keyboard with vowel omission enabled.

I found the opposite when I went through a split keyboard phase and did Colemak-DH layout. As far as english goes I found that a lot of words had a smoother "flow" when it comes to finger location.

In the end I went back to a regular qwerty because my WPM on split keyboards/colemak-DH was considerably worse even after many months of practice.

I can believe it. For a physical keyboard, I would much prefer a layout with a DVORAK-like home row, but probably not for mobile. The imprecision of touch and swiping text entry likely do markedly worse when high frequency characters are on top of each other.

I'm sure somebody has created a software keyboard where the more used characters are simply larger than the less used ones.

Pretty sure Micheal Phelps has fins though

Even at a bit less than 200WPM, my physical ability to press keys is seldom the limiting factor for composing prose, let alone for coding. I would suspect that this is the case for most halfway competent keyboard users using more than three fingers, and that QWERTY is “good enough”.

  The keyboard arrangement was incidentally changed into QWERTY, first to receive telegraphs, then to thrash out a compromise between inventors and producers, and at last to evade old patents.

Interesting article. The connection to Morse code makes a lot of sense (C being similar to S). The requirement to move I below 8 to type 1870 or 1871 quickly is hilarious in retrospect. At the time who could have known the decision to focus on efficiency for the coming decade could be so enduring?

I don't follow the connection to Morse. Can someone summarize their argument in a comprehensive way?

The first typewriters were for telegraph operators turning morse code into written letters.

In Morse telegraphs, some words are abbreviated and so characters have a different frequency distribution. For example RECEIVED may be shortened to RX, making X much more frequent than otherwise.

Very interesting article. I don’t understand, however, how shorthanders used typewriters for short-writing. The figure on p.168 (above fig. 9) is not explanative.

The numbers above the words indicate which finger (index, middle, ring) is used to press a key, the letters below indicate the hand (left or right). Basically a precursor of touch typing that doesn't use the little fingers and doesn't always use the same finger for the same key.

The actual shorthand would be written on paper, with the typewriter being used to expand it to a more readable form.

Thank you. I mistakenly thought the typewriter was used to type shorthands.

Now why was fricken Semicolon given the prime real-estate of being an unshifted key?

Because the smart way (with "M" there instead) was patented. From TFA (emphasis mine):

> In order to evade the patents that were assigned to the Type Writer Company, WS&B slightly changed the design of No. 2 including the keyboard arrangement (Fig. 9), where M was moved next to N, and C was exchanged with X. It was the QWERTY keyboard arrangement as seen nowadays.

If we consider 100-year-old prose, we used the semicolon in places you see a dash now; like this - not this. It was important; it was located with period and comma and apostrophe.

To make it easier to code in C and Pascal ;-)