Saturday, July 28, 2012

On the Design of Simple Things


The Design of Simple Things

by Jeffrey Chan

On simple things

In the pursuit of complex design, we have neglected the much harder problem of designing ‘simple things’.  
By ‘simple things’ I am not invoking Don Norman’s concept of “everyday things”--practical things that we interact on a practical, everyday basis. Many of our ‘everyday things’--plasma TVs, iPhones, and even mechanical watches--are hardly simple things. And by ‘simple things’, I also do not refer to ‘minimalistic things’. ‘Minimalistic thing’ expresses a deliberate art of irreducibility, but a ‘simple thing’ is the art of straightforwardness. An iPhone is minimal but not simple. In contrast, a man’s wallet is straightforward but hardly irreducible: there can be multiple variations in materiality and configuration based on the same design. Finally by ‘simple things’, I am not implying ‘simple-mindedness in designing things’, since it is possible to be straightforward without being simple-minded. 
What then are ‘simple things’? Is ‘simple things’ a category for things or is it a quality or attribute of things? By ‘simple things’, I refer first to a category of design artifacts that do not entail compound or complex operations for the user. Yet at the same time, ‘simple things’ also tend to embody the design ideal of a final and stable design solution: it cannot be made ‘simpler’. For this reason, I also contend that any design of ‘simple things’ can reveal to us clearly some of the most fundamental challenges of design, since there is very little room for extraneousness or irrelevancies. For example, it has been said that a pair of scissors is one of the successfully resolved ‘simple things’: its basic morphology is final and its design configuration stable because it is the clearest design statement--up to now--for accomplishing single-handed shearing. 
Seeking final and stable design solutions in every category of design is therefore a very attractive proposition. Final and stable design solutions possess the threefold virtues of clarity, predictability and stability. Because of this, they tend to minimize any undesired and unanticipated consequences, barring human error or willfulness, that usually accompany untested and evolving design solutions. To suggest that final and stable design solution is the teleology of design is therefore not an exaggeration. In design, we usually begin from a stage of disarrayed complexities to a stage of synthesized simplicity, where every single element is ordered to everything else; and which inadvertently is always an arduous process of winnowing by accretion. Whether such final and stable design solutions once attained might lead us to a arrested state of permanent and existential design boredom is debatable; but the dividends that can come with these final and stable design solutions are however indisputable. 
But designers today tend to avoid discussing and the making of ‘simple things’. In place of ‘simple things’, we have a plethora of overly complicated designs reflective of our seemingly overly complicated lifestyles. As suggested, ‘simple things’ are not ‘minimalistic things’. Yet today designers prefer to take the minimal for the simple, preferring an appearance of simplicity rather than the quality of simplicity. It is plain to see why ‘simple things’ are not popular in contemporary design: they neither present complex features that invite our curiosity nor do they seduce with alluring forms. They are, in short, icons of banality. After all, what is left, except for infinite styling and mannerism, when final and stable solutions embodied by ‘simple things’ have been discovered? Moreover, the interest-triangle between the industrious designer, the industry and the cultural society continues to demand for more exciting variety rather than successive depths of sustained profundity in design, resulting in the moral hazard of producing ever more superficialities masquerading as the novel and the experimental, something Susan George rightly calls things that make the status quo look brand new.  
* * *

Christopher Alexander’s example on the design of a simple thing

Christopher Alexander may not be the greatest architect and designer alive today, but I suppose he may just be the greatest designer capable of feats of design introspection alive today. We may be willing to fault Alexander for his metaphorical and sometimes mystical language; but we should be unwilling to dismiss the substantive content that he tries to convey through this language. And like the philosophers Heidegger and Wittgenstein who each cycled through polemical phases in their career--for Heidegger from early mysticism to later rationality and for Wittgenstein precisely in the opposite direction--Alexander began his career as a design rationalist but who later disavowed the rational approach for a more mystical (i.e., read ‘inexplicable) turn towards design and design thinking. But unlike philosophers who could make a clean break with their prior philosophical frames because they transact mainly in fluid ideas, for an architect or a designer, this break is much harder to achieve, since the medium of design--its processes, ideas, materialities and outcomes--must invariably be located in the empirical world and therefore a constant despite ideational transformation. For this reason, Alexander’s contemporary work still offers a wealth of design insights accreted from the best of his past experience as a rigorous design rationalist on top of his ongoing explorations as a design Platonist. 
In one of his more recent writings, Alexander asks us to contemplate on the design of a simple Japanese chisel. Here, it is worthwhile to cite his passage almost in full:
“Consider a simple Japanese chisel. It has a handle, shaft, blade, tip, and so on. It works because of its centers. Each kind of working that exists in the chisel inheres in one of the centers which we see in the chisel. The tip of the blade is the cutting edge--this is a center that does the cutting...
The tip of the handle is bound with a steel ring, so that if you hit it with a hammer it doesn’t split. The steel binding at the lower end is another center, which stops the wood from splitting. The shaft of the blade, where the blade meets the handle, allows a handle to be out on the piece of steel, and makes a firm connection. This shaft is another center, and it does the work of making a firm connection between handle and blade...In a well-made chisel, the geometric centers correspond exactly to the centers of action (i.e., the way the chisel interacts with the world when we use it, how it cuts the wood, holds in the hand, gets hit by the mallet, and so on. 
The functional life of the chisel, the excellence of a good chisel, does not come merely from the fact that these centers exist. It also comes, as in the case of the little stars on the tiled wall in the Alhambra, from the way these centers help each other. For example, the shaft is tapered one way and the wood alongside the shaft is tapered the other way. The shaft makes the wood more of a center: the wood alongside makes the shaft more of a center. The binding strengthens the existence of the blade by being at the other end of the blade. The blade strengthens the existence of the binding as a center.” [from Alexander, C. (2002). The Nature of Order, pp. 406]. 
While Alexander did not specifically allude to the chisel being a ‘simple thing’ or to any specific discussion on final and stable design solutions, it is however clear that the “good chisel” conforms to the twin criteria of final form and stable design configuration. In other words while it may be possible to modify or even enhance the weighting of the different parts for different users, it is however not very productive to re-form the chisel, because the design of this chisel with a series of interdependent centers has been discovered and perhaps even, finalized. But what is most valuable in Alexander’s writing is his ability to move from the material world to a level of design thinking that can be applied to just about every kind and scale of design activity. Perhaps the only caveat here is that Alexanders lessons do not, and should not, be applied to design that must result in the deliberate de-centering of different functions, for example, in Chuck Perrow’s preference for a more loosely-coupled systems for the design of a nuclear power plant, such that a small error will never concatenate into a catastrophic failure. 
* * * 

On the design dilemmas of a ‘simple thing’: designing an ideal water dispenser

The question that Alexander’s case then raised is: to what extent does his theory of interdependent centers in design apply? My answer: it does not go far enough. While the general lesson by Alexander should apply to the design of many things, it however does not apply to the design of all ‘simple things’. In my case for example, to the design of a (simple) water dispenser. Furthermore, my case will attempt to demonstrate the opposite lesson from Alexander’s: that centers can nullify each other. I contend that this latter possibility of nullifying centers may perhaps be as important for design epistemology as centers that are mutually reinforcing.
I introduce my case through two existing water dispensers available in the market today. The first one is called ‘Ovobar’ and the second is called ‘Penang’ (simply because we purchased it in Penang). I want to emphasize that these two water dispensers are perfectly workable designs for dispensing water, although one is geometrically and certainly more pleasing than the other (guess which one). 

(photos: left, Ovobar, source: aquaovo; right, Penang, source: author’s own)
As a quick overview, both do not work very well for reasons to be explicated. After all, for most of our recorded history, we have been accustomed to drawing water upwards (from a well). In contrast, dispensing water downwards (from a dispenser) is an activity that we are apparently still grappling with in design. And as a matter of due diligence, I have interacted with the Ovobar (rather extensively I must add) in Venice, California, and Penang is an everyday fixture in my kitchen. My aim here is to extract a set of issues from these two designs that can represent my case of nullifying centers in designing ‘simple things’. 
So what are the design parameters and goals for an ideal water dispenser? I will answer this question based on my own experience. 
First, a water dispenser collects water from either the faucet or a water filtration system. Instead of either turning on the faucet or the filtration system whenever one needs a drink--and in tropical Singapore that is quite often--the water dispenser performs this role in their stead, thereby extending the longevity of the mechanical faucet and the electrical filtration system. 
Second, a water dispenser must have at least a minimal volume. By minimal volume I mean a volume capable of supplying drinking water to a family of four for at least one full day before refilling. Conversely, there must also be a limit to how large this volume is, before it becomes unwieldy such that it exceeds the normative carrying capacity of a single person who has to haul it from box to kitchen counter.  
Third, flow dynamics and other efficiency issues. By flow dynamics, a water dispenser will only dispense water if a volume of water exists above the dispensing tap. Ideally, this tap should be positioned as low as possible for the greatest amount of pressure (based on the most number of uses) and for as little water as possible to remain unused (i.e., water that remains below tap level). 
Fourth, the water dispenser must be able to dispense water to a wide variety of different vessels, for example, cups, tall glasses, water bottles, bowls, etc. The water dispenser must accommodate for all these varying geometries and sizes as much as possible. 
Fifth, the water dispenser must be stable. This seems sufficiently clear: no one wants to mop the floor or pick up the smashed pieces of a toppled water bomb weighing on the average 25-30kg in weight. 
Finally, the water dispenser should be made of a material that is easy to clean and maintain. The geometry should facilitate maintenance, and the form of the water dispenser should be pleasing and fitting in a wide variety of different kitchen designs. 
Clearly by the goals cited, the ideal water dispenser is one that meets all six design goals listed above. Conversely, the most undesirable water dispenser would be one that does not meet any of these six goals. 
Now there can be many more design parameters and goals for the ideal water dispenser. But the listed six are what I think the major and vital ones. Even so, a designer trying to meet all six major goals will inadvertently run into trouble, assuming that the designer continues to work within the frame of a prototypical water dispenser. 
For a start, the major ‘tri-lemma’ in the design of a water dispenser is between the position of the dispensing tap, the attainment of maximum water pressure (for the most number of uses) and the maximum volume of water that can be dispensed. The obvious solution is of course to position the dispensing tap as low as possible (towards the level surface). But this obvious choice, even when it maximizes water pressure and also volume of water that can be dispensed, unfortunately contradicts the fourth design goal, which is to accommodate as many vessels of different geometry and size as possible. A design compromise between all three goals would be to position the tap somewhere above the height of a normative teacup, which is what Penang has elected to do. As brilliant as this compromise was, it however always leaves a large volume of water below the tap level while forcing the users from using taller cups, water bottles or anything that exceeds the height of a teacup. To circumvent this tri-lemma, Penang has elected to prioritize two of the three design goals--for maximum pressure and maximum volume over maximal accommodation--to force a design compromise. For to position the height of the tap higher (i.e., prioritizing goal number four to maximize accommodation for different vessels) would immediately compromise the perceivably more important goals for maximum pressure and volume. On this, it is perfectly reasonable to think that design compromises and designing reinforcing centers are two viable, but distinctively opposite, modes of functional design epistemology. 
However Ovobar took a different approach. Rather than to rely on the utilitarian approach of Penang, Ovobar attempts to circumvent the tri-lemma through the introduction of a stand. What the stand does is that it effectively preserves the design goals of maximum pressure and maximum volume (i.e., positioning the tap to as low as it reasonably should) while accommodating to a wide range of drinking vessels of different geometry and sizes. For the ‘chic’ modern kitchen that the Ovobar is envisioned to operate in, it seems reasonable that it should at least accommodate tall stemware, which is precisely what the stand is for (other than for the obvious stabilization that the egg geometry requires...).
But Ovobar neglects an emergent design consideration that implicates any elevated design. The elevation of the water dispenser on a stand with limited (contact) surface area immediately increases its instability (relative to one that is positioned on a level surface). Moreover, as water is dispensed, the dispenser becomes lighter and weighs less, thereby rendering it even more unstable with a diminishing base weight. Since the overall weight of an average water dispenser is primarily attributed to the weight of its carried content, this is an important consideration that should not be overlooked. On this, Penang has elected to go for a stodgy but passively stable design; its profile flares outwards to form of a wide and solidly thick base that attempts to decouple (or at least reduces the moment) for the dispenser because of its differential weight through use. 
* * * 

Centers? Nullifying centers for Reinforcing Centers

The water dispenser is undeniably one of these ‘simple things’. Yet as depicted, the design of one is anything but ‘simple’. The position of the tap is clearly a chief center for the design. But any move to reinforce this center with other centers (e.g., design features for maximum pressure or maximum volume of water that can be used) nullifies another center, which is the accommodative design for this water dispenser. To circumvent this, a stand was introduced (which is a new center). But this new center introduced in the hope of reinforcing all other three centers create its own contradiction, which ironically makes the water dispenser with stand as a complete system increasingly less stable and more precarious with use.
But at least a few lessons are clear. First, certain centers work together with other centers; they are either pairs that are integrally tied together or pairs that can be amplified by being tied together. For example at least in the water dispenser, maximum pressure and maximum volume of water that can be used are integrally paired centers: the more one increases the more the other increases as well. Second, there are overruling or chief centers, or a center that affects all centers. The tap is one such centers, such that any move on this center preponderantly affects the rest. Finally, new centers can be introduced, though not without introducing a whole new set of issues that would then subsequently recaliberates all existing centers. 
If Alexander’s example of a chisel is startlingly simple, it is also a startling over-omission on quite a few other issues that can occur in the design of a ‘simple thing’. What my own example attempts to demonstrate is that the designer must also invent new concepts to cope with the possibility of nullifying centers. What are the pairing centers? How do we identify them? What are the chief centers? What are the strong centers and what are the weaker centers? And when do I break the existing morphology (or form) to introduce new centers that can recaliberate or circumvent the limitations of existing relationships between centers? This case of nullifying centers hardly overturns Alexander’s case for the mutually reinforcing centers; rather, it extends the cognition necessary for the design of a successful ‘simple thing’ in recognizing that mutually nullifying centers exist as well.