Archy

I still find a lot of inspiration from the earlier UI books (ideas), before everything settled down. Lots of crazy, awesome ideas. Now it's just how to books.

I've been anticipating the ZUI for ages. eg It's the correct answer for a window manager. It's the only way to maintain continuity, preventing users getting "hyperlost", kinda like what breadcrumbs try to accomplish.

Did you see the new flashy zooming Calendar transitions in the WWDC Mavericks demo? It appears to have the right balance between zooming and level of detail (LOD). Me want!

FWIW, I did not anticipate the dynamic 2.5D in the forthcoming iOS 7. That's gonna be huge.

I've thought it was archy as in anarchy, and until I clicked I was wondering why someone would post article about some obscure political term.

Fond memories? Archie was truly horrible.

The Active Oberon System (AOS) use the zoom-approach. There are some videos in youtube. https://www.youtube.com/watch?v=Pe0ZdzO_urU

The 'zoom approach' is used in inkscape and adobe illustrator.

There is no need to remember anything actually. If you've ever used Enso, Ubiquity (both made by Jef's son Aza) or Sublime Text 2, or Google Instant/autocomplete, or indeed Firefox's "Awesome Bar", then that is the kind of experience the humane interface describes.

Menu and typing commands are not either-or exclusive; you can have both. In a system like Archy there can be menus stored as plain text - given that it's trivial to select text and invoke it as a command. It's just that menus don't require a separate subsystem or component, they're stored in the homogeneous system's data representation (and they can be user-editable, which modern GUIs don't allow - yet).

I've known of Archy and The Humane Interface for about ten years, and it's funny how more and more interfaces are converging towards the interaction style defined by it (with some tweaks because of backwards compatibility and-or because some of the ideas in the book have been tested and improved since then).

I wrote a ZUI at my previous job. We needed to speed up the process of manual classification for large number of products. The web based approach did work but we figured there might be a better way. The idea was to split large number of product into smaller sets using various tools. These small sets were then manually classified (with suggestions provided by the system).

All this was done in a zoomable interface, so it was very easy to navigate the sets. The system was a huge success.

Here is a short video. Sorry about the quality: https://www.youtube.com/watch?v=CLy6Qj7MNOo

Most people who aren't used to (decent) tab completion interfaces have this misconception. Let me guess: you've never really spent much time in a CLI like bash or Emacs? It's incredibly powerful to try tabbing into what might be related. Emacs tab completion (variables, functions, buffer names, etc) in particular is very nice in that you can type part of what you're looking for and anything that matches in any part will show up in the completion list (eg M-x -mode<TAB> will get you a completion list of anything containing "-mode" anywhere in the function name).

As Steve Yegge points out in Effective Emacs (https://sites.google.com/site/steveyegge2/effective-emacs), mousing through a menu is very inefficient, even when you're learning or exploring. They don't scale well, can't possibly cover all the options, and once you've dug down two or three deep, you're sort of trapped. Whereas with decent tab completion, you can backward-kill or just plain <DEL> your way back up the heierarchy.

FUNNY BECAUSE I DON'T HAVE A PROBLEM HOLDING SHIFT WITH MY LITTLE FINGER.

I used it. It was a very interesting experience, but a little hard to get used to. The position of alt is not comfortable for the quasi-mode required. CAPS is more accessible, but then you can't use your pinkie. Raskin's idea was to use a special keyboard with two spacial keys, LEAP FORWARD and LEAP BACK near SPACE.

It's so sad that the project died. The interface was very promising and I long for that keyboard I never used.

I've tried it, and it's a pain even with dedicated keys.

This is one point of Raskin's design where the rationale is sound, but the solution he found was sub-optimal. Instead of a "pure" quasimode, modern interfaces do incremental searching following the search box style introduced by Firefox, with a persistent dedicated and non-obtrusive search field.

I find that this is a superior interaction, and it still can be considered non-modal. The style has spread to notifications too, with browsers showing small rectangles above and below the window instead of an obtrusive modal alert dialog.

The LEAP keys are supposed to be under the space bar, so that they are easily held down while your fingers are in the home position. Try it, it's surprisingly comfortable.

I would pay good money (a few hundred dollars, easily) for a mechanical keyboard that had those two additional keys under the space bar, as well as a vertical F-key cluster left of the main typing area, as seen on old IBM PC/AT keyboards.

A big difference between Emacs and Raskins' design is the concept of modelessness. If you've ever typed META+X in emacs in anticipation of typing a command, and then switched windows or got distracted or clicked the mouse, emacs is left in the 'command mode' and will misinterpret your next input.

Raskin coined the term 'pseudo-mode'. The difference between a pseudo-mode and a mode is the difference between the SHIFT key and the CAPS-LOCK key (for entering uppercase letters). The latter frustrates while the former goes unnoticed because it works so well.

If you are in any way involved with usability, read The Humane Interface: http://en.wikipedia.org/wiki/The_Humane_Interface

Those are details that can be refined later. The overarching concept is great.

IIRC this is the guy who pushed for graphical, user-configurable types in the original Mac - something that hadn't been done before ever. The software was an alpha release, it's likely that they went with whichever fonts were available in the environment they used.