Englebart's Violin (2012) | Hacker News Reader

Englebart's Violin (2012)

loper-os.org

53 pointsDanBC20 comments

I got really excited about chorded keyboards and did some research in the past.

The chorded keyboard is not as efficient as two handed QWERTY. Englebart reported getting 35 words per minute with his right hand and 25 WPM on the left after extensive training[1].

Granted, this is a good speed for having one hand free to use another type of input (pen, mouse, touch screen or whatever).

A chorded keyboard is also a promising idea for mobile computing where a dedicated physical qwerty keyboard is unwieldy.

I wasn't able to see enough of a benefit to try building one or purchasing some of the existing products. At the right price point, I definitely would try it as a novelty.

It seems there is a chicken and egg problem of inputs and software that makes you efficient using them.

[1] http://www.billbuxton.com/input06.ChordKeyboards.pdf

Englebart's chorded keyboard entered a single character for each chord. Most chorded keyboard systems type multiple characters or entire words in a single chord, which is how you get people doing 3-digit WPMs, and how court reporters and captioners on Stenotypes can keep up with speaking.

Of course, if you wanted to do programming on a chorded keyboard the chords you need are totally different, but apparently it's still very doable: http://plover.stenoknight.com/2010/04/writing-and-coding-wit...

It's interesting to see 35WPM for that. I read somewhere about half-QWERTY with one hand up to the 65WPM range.

While it's interesting to look into chorded keyboards specifically, the take-away for me here was more the general idea that it's not terribly difficult to come up with interfaces that are objectively more effective than the status quo by a pretty serious margin and still completely fail to achieve large-scale uptake.

On the other hand, I believe there's lots of room for improving on the status quo in a way that isn't necessarily intended to ever supplant the conventional alternatives, just focusing on some subset of users who have a good reason to be willing to invest a little more time in becoming more effective. (This is part of the process that led me to design my own custom keyboard.) If you look at Emacs or Vim users, it's clear that some people are willing to spend a lot of time learning skills that improve their efficiency.

You need to do some more research. Stenotypes can write faster than people can talk. There are studies of expert Twiddler users who can write at closer to 100wpm.

I really wish someone would come out with a nice, not-too-expensive-but-not-too-shitty USB steno keyboard.

Check http://www.openstenoproject.org/ - they have some pretty great ideas and alternatives (and the software to run them)

There is a type of stenographic keyboard which is common in Italy and, with minor modifications, I believe you could re-purpose a MIDI keyboard for it. The keyboard is called "Macchina Michela" and the associated method is "Metodo Michela", you can see it in action in this video (the audio commentary is in Italian): https://www.youtube.com/watch?v=uniHoxe6HpY

Stenosaurus is apparently making some progress: http://stenosaurus.blogspot.ie/ .

They use abbreviations that can be ambiguous. Not really the same.

Velotype/Veyboard is apparently more regular http://www.velotype.com/en/ http://www.veyboard.nl/en_main.html , probably more suited for general use.

One part of me thinks it's terribly romantic to envision a world where the professional gets to use the big, grand tools to practice his craft, but I think it's actually more comforting to know that anyone with access to the most basic tools can create. It might seem like we're all forced down to the lowest common denominator, but I think the playing field might be better for it.

The article mentions Emacs (and I would add vim) and throughout I could only think about these two. They are the tools that reward learning and multiply your efficiency. It doesn't all have to exist in hardware.

You can always get one of these(0).

0 - http://www.dx.com/p/fs3-p-tri-pedal-usb-foot-switch-controll...

Visual Studio. The grand tool the author is thinking of is Visual Studio. It is light years ahead of anything vim and emacs can do, though I concede that basic text entry and editing is slightly more clunky in it. Then, at the really high end, you have tools like Intel's VTune which let you profile your application in ways you never could with only Visual Studio.

The author's gripe is pretty much that professional tools have free versions for the home tinkerer and affordable licenses for the lone professional.

> You can always get one of these(0).

I did. Sadly, they suck :(. First of all, there's a problem getting firmware that would run on Windows. Secondly, getting the pedals to simulate raw control keys is tricky (but can be done). But thirdly, the construction isn't too solid - I had them in the office for a while, and somehow someone (I suspect the cleaning crew) managed to break off one of the pedals. :/.

I smell a market ripe for the taking... Or maybe not.

With that said, I actually thought about buying one of these, but I ended up just revamping some of my Emacs bindings. Then I ended up switching to vim, so chording became less of an issue in general.

(When I still found myself wanting the Emacs operating system I went with Spacemacs. It really is the best of both worlds.)

I suppose this is an eloquent defense of APL.

Thanks for posting this 'DanBC, this is golden! It captures exactly the issues I sometimes rant about in UI/UX threads on HN - that the IT industry is focused on producing toys, not tools.

The article focuses on hardware, but its point is more general, it touches software too - basically the solutions developed both "by developers for developers" and those created for non-tech users. There are lots of things that worry/sadden me about this.

Within the industry, I'm tired of hearing comments like "don't use advanced feature X, it's hard for the junior programmers". Like, "don't use lambdas / streams in Java 8; it may be more readable for you, but it isn't for the cow-orkers who don't know lambdas". Well, one would think that it is expected of a professional to occasionally learn some stuff. But what I see instead is "best practices" aimed at the lowest common denominator creating a culture of code monkeys.

As for end-user software, I'm tired of ranting about it, and - as I learned from this post - Erik Naggum (like usual) does a much better way of explaining the problem. So I'll just end with a graph.

  power                                                                    /--
    ^                                                                 /----   professional
    |                                                             /---          software
    |                                                          ---
    |                                                         /
    |                                                        /
    |                                                       /
    |                                                      /
    |                                                     /
    |                                                    /
    |                                                   /
    |                                                  /
    |                                                 /
    |                                               //
    |                                             /-
    |                                           /-
    |                                        /--
    |                                      /-                             "user-friendly"
    |                                    /-                                  software
    |                                  /-                         /---------------------
    |                       /-------------------------------------
    +-----------------------      /-
    |                        /----
    |                 /------
    |         /-------
    |  /------
    |--
    +--------------+-------------------------------------------------------------------->
                   |                                                                 time
                   |
                   |
         point where companies make their sale
    and beyond which they don't give a shit anymore

We could be doing so much more with the power of computing if only we gave up on the expectations that people have to master of their software tools in 15 seconds since first seeing them...

A similar issue is for the commodity hardware handed out to developers. The desktop setup gets standardized to the lowest common denominator, managers/IT do not get the argument that saving 1k on a developer's desk setup is not good business practice.

I get the same kind of hardware as a developer/researcher than a secretary, some low-end lenovo laptop with a single standard sized screen, running the same IT-provided windows image with performance-crippling endpoint and scanner software. All my development happens through a VM which connects to a remote server, via ssh/emacs tramp. Of course this invites the argument 'you are using your desktop as a terminal anyway, why do you need something more powerful'.

Engelbart, not Englebart

I replied to this: http://www.loper-os.org/?p=861&cpage=1#comment-16138

Reply on news.ycombinator.com