notes

Finding baseline for users. Disney’s early tribulations adding sound to film (prior to dubbing) included the use of an onscreen bouncing ball. A visual metronome for orchestras and sound engineers. The mnemonic translates the (foreign) language of animation to the (familiar) language of music. When musicians are the users, the baseline is a beat.

Interface as a key. Disabling the steering wheel (usability, simplicity, automation) as security. The interface as removable layer without which the machine is cognitively and/or ergonomically too difficult.

With digital UIs, it’s easier to create alternative UIs—workarounds if the standard UI is removed (or otherwise poor). Actually, what’s easier is the ability to share and proliferate a UI hack. Manufacturing makes it harder to share physical UIs (ex. old cable black box)… except when it’s not (ex. a BIC pen to unlock a bike U-Lock).

Good fences, good neighbors. Transparency. Gated communities.

Audio messages, an afterthought in the US, are more popular in Morocco. Voice functionality makes WhatsApp more accessible—technically and socially.

Typing acumen. Typing on a touchscreen can, even for experienced users, be problematic and frustrating. Add to this that many Moroccan users have spent far less time with a physical keyboard—an important intermediate step on the learning curve towards the mobile touchscreen. Add to this that, if you intend to type in Arabic, you may have to input Arabic script via a French or English keyboard (languages you may not speak). Voice messages are easier.

Like many North African cities, Fes is built around the medina—the original city, walled-in hundreds of years ago for protection. Inside, the alleyways are narrow (I’ve read the minimum is dictated by the width of a loaded donkey, but even that seems generous in particular corridors). In addition to laying claim to the title of oldest city in the world, Fes stakes itself as the largest urban area without cars. The “last mile” here is substantively different.

The medina gates (the Babs) serve, naturally, as transport hubs. Carparks here are key landmarks for providing directions; the bus, the taxi and it’s many cousins queue as well. For people, the last mile is by foot. For goods, the last mile is varied—donkeys, hand-drawn and semi-automatic carts gather to carry goods inside the walls. The need to transition, say, crates of onions from a van to a donkey is a filter on the restaurants and stalls receiving them. The transition regulates the processes governing the size and type of commercial activity taking place inside the walls.

Supermarkets and other large commercial ventures are absent not for lack of demand, but lack of supply. There’s no means to feed beasts of that scale. The 150Dh per day rent of a donkey might be feasible, but at a max of 100kg per trip (about three boxes), there’s no efficiency.

The lack of cars isolates the medina preserving a space for the small shops and artisans to remain in place. Industries ousted by Western expansion can remain lively here—the tanneries, the metal- and woodworkers, the milkman.

There are other consequences of the absence of cars. The crampedness of the city is evident. Scaffolding everywhere allows construction atop construction. Beams above the alleys keep buildings from collapsing into one another.

Labeled nuance (when the offer is sophistication, complexity, craft).

Machines develop around the shape of their human operators—literally, in physical form and figuratively in cognitive operation. The object often just has to fit in the right ways. The handle of a vacuum needs to match the hand (well enough) and it’s height needs to match human scale (well enough), but it doesn’t need to mold much to a toe (except to stay out of its way). The interaction between machine and user is reflected in this fit.

Loose physical fit can be the result of tighter cognitive or emotional fit. The interior of a car is very much shaped and scaled according to human proportions, but it doesn’t “fit like a glove.” The buffer space serves other purposes, including comfort, luxuriousness, interaction between passengers, and general flexibility in the use of the interior. Perfect fit should, therefore, rarely be expected to manifest as perfect physical fit. 

Hap tip: Henry Dreyfuss.

When a person uses a machine, she closes the gap between herself and the object. That is, she operates a machine by reaching out, bridging the buffer space, and creating a point of contact with the machine. Because steering wheels aren’t attached directly to the body, drivers must extend their arms to manipulate them.

When fit is loose, it’s common to improvise solutions to fill the space. Hacks may improve physical ergonomic fit, but can also serve non-physical functions. Users may add Post-It reminders or button labels to aid memory. They may install a curtain over a window for privacy. They might also do things to improve other types of cognition cycles, for example, to create better fits with their daily routines. Storing a toothbrush in the bathroom is more convenient than storing it in a desk drawer away from a sink. These are features that aren’t built into the original design. Hacks improve technology by tightening fit. In doing so, more processes are automated, fewer functions are performed by the user.

Hacks are at the same time component and machine. Though they come into being in response to existing machines and may be less useful without the prior, they’re not necessarily dependent on the originals. They can be independent add-ons, reusable in other places, or incorporated directly into the original machine. A cupholder built in to a car dashboard may be more stable (physical fit) and less tacky (cognitive fit) whereas a detached one can be installed on either the driver or passenger side or in an altogether different vehicle. Hacks narrow the gap in both cases. When incorporated into the original machine, they remove the need for the user to manage the hack as a separate entity (i.e. the user no longer needs to mount the hack) while also removing a degree of modularity (i.e. the hack is now hardwired and can not be removed or re-purposed elsewhere).

As the gap between user and machine closes, humans are progressively freed (or displaced) from operating machines. Hacks extend the machine, then progress to such a degree that they oust the user—fit is tightened and then some. In replicating human operation, hacks are robots. At introduction, new technology has to work with and within existing infrastructure thus robots, at least initially, take the form of earlier machine designs. The switches and levers of the prior machines, having been cast around the mold of human anatomy, now turn around to form a mold from which robots are cast. Because prior machines were shaped around humans, robots resemble humans.

Jean Baudrillard:

Because the automated object 'works by itself', its resemblance to the autonomous human being is unmistakable, and the fascination thus created carries the day. We are in the presence of a new anthropomorphism. Formerly the image of man was clearly imprinted in the morphology and the manner of use of tools, of furniture, or of the house itself. [5] In the perfected technical object this compliance has been destroyed, but it has been replaced by a symbolism of superstructural rather than primary functions: it is no longer his gestures, his energy, his needs and the image of this body that man projects in to automated objects, but instead the autonomy of his consciousness, his power of control, his own individual nature, his personhood.  

In closing the gap, the design is less constrained to the human form. It only needs to incorporate the human in whatever respect the human is still involved. The Roomba, in replacing the traditional push vacuum, for example, is operated by a button on the top surface and only needs to adapt to the human on this level. It’s shape is more dramatically dictated by it’s interface with the environment and by it’s cognitive fit with the user.

Last week’s brain dump on 3D Touch was based on the Keynote introduction rather than firsthand experience. Now with all of 24 hours to test it out, here are additional thoughts, written with the caveat that grasp of force interaction is likely to improve with more time.

1. Peek and pop are inconsistent with the rest of the multitouch terminology. Pinch and swipe describe physical human gestures, as do tap, drag, and swipe. Peek and pop describe what the device is showing, or maybe they describe the user’s goal. They don’t describe what your fingers are up to. Something like light and firm press (which would make press distinct from tap) makes more sense.

2. Peeking and popping essentially provide a better (in my opinion) way to present and dismiss preview modals. The interaction is better, but modals themselves present mirror-in-mirror issues and complicate the interface model. Example: From my inbox, I peek at an email. The email preview shows a date underlined. If I were in the email proper, I could peek at the date, but because I’m within the preview modal, when I try to peek, nothing happens. At the moment, I don’t think any interaction is allowed within the preview. Wisely, there is no peeking inside a peek, no smaller modal within the modal. That makes sense, except that the result requires the user to remember when peeking is allowed and, whether viewing properly or previewing it, the email looks the same. Sure, there’s a modal frame, but this isn’t enough to prevent the instinct to tap blue underlined text.

One possibility is to gray the previews in some way—use grayscale, light transparency, a slight blur effect, etc. Make it obvious that they’re disabled and there’s no interaction with the preview. Make it more obvious which mode you’re in. This makes the preview less nice and less accurate, but it is a preview after all. In theory, it should be used for shallow browsing. Not sure I love this solution.

3. Pop has the same effect as a normal tap, right? It just allows the user to go straight from peeking into a detail screen. This seems unnecessary. The model seems more streamlined if the user just releases the preview and taps.

Force provides much greater input resolution than a plain tap. Because the gesture is read as a continuous range, it seems it should, where possible, be mapped to output that’s also continuous. At the moment, force is broken into discrete chunks—force below a certain value triggers peek, force over that triggers pop—with no steps in between. In the current Photos app, lightly pressing an image shows a peek preview, a firm press pops open the image properly. Pressing firmer could, instead, progressively increase the size of the preview. With the dimensions tied to force, the user would see and control the transition. In this case, force is entirely about previewing and when pressing stops, previewing stops.