Abstract Design Situations - Design isn't a Shape and it can't have a Centre

Willy the web designer decides to place the main navigation menu on the left side of a web page. He decides on the names of the navigation links and their order. He next decides to test out this menu design with a range of colleagues spanning a range of job roles. He prints out a wireframe and ask colleagues about their expectations about each link: what would selecting this link do?

Willy has made several choices here, but are they the same type of choice, or is there something that distinguishes choices about a web page menu from choice about its evaluation? There do appear to be two types of distinct choice here: one concerns choices about a design, the other concerns choices about its evaluation. The former choices are made within a craft space defined by the possible layout and content of a web page. The latter choices are made within an investigative human space defined by a range of evaluation practices, scoped out in part by choices of evaluation tasks, participants and materials (i.e., paper). We can thus regard choices about designed artefacts and choices about design evaluations as distinct. There is no overlap between the craft space of web page layout and context, on the one hand, and the investigative space of evaluation study design on the other.

Even though the above is most probably blindingly obvious, it draws attention to different types of design choice.  Many people may tend to think of design choices as only relating to designed artefacts, when in fact they relate to a wide range of design activities, such as evaluation.

Willy chose to use an opportunity sample for hallway testing. One of the participants is an accessibility expert. She immediately points out that the menu contains many links, which could cause difficulties for visually impaired site visitors who use a screen reader. Willy asks her to read the menu, which is for a van hire web site.  He argues that few visually impaired users would be using a van hire website, and that the broad shallow navigation structure enabled by a large top level menu would favour a range of site visitors, especially regular users who would appreciate short click paths.  Here (rightly or wrongly), Willy has made a further type of design choice, in this case, a choice of beneficiaries. He has decided to give limited consideration to the needs of visually impaired users, preferring to optimise the site for sighted regular customers.

Another participant is a security guard, who was happy to spend time discussing van hire web-sites at great length. He told the story of a friend who had great difficulty finding the depot to pick up a van from. He was delayed and had to rush to complete his pick up and delivery and return the van to the depot. The van was too short for his load, and he had to tie the rear doors together with his trouser belt. Willy realised how stressfull this must have been, which got him thinking about the capabilities of the web site. If he had good directions for getting to each depot, and clear indications of the loads that each type of van could carry, then this should avoid the poor experience of the security guard's friend. We could regard these choices as ones about the designed artefact, and indeed they are, but they are very different to the location, order and content of a navigation menu. Willy has begun to reflect on the purpose of the web-site: is the purpose simply to provide features for hiring a van, or should it do all it can to ensure a comfortable driver experience from getting to the depot to returning the van?

There are pros and cons of treating design purpose as a distinct type of design choice from choices about artefact features. I feel strongly that it is a distinct type of choice, as it can be made without reference to any details at all of web site features.  However, even if it is not, we can see that there are different types of design choice.

I have developed the concept of Abstract Design Situations (ADS) to move beyond normative one-size-fits-all constructions of design, e.g., user-centred, value-sensitive, human-centred, universal, or sustainable. Such positions on design tend to be formulted for specific craft practices, e.g., interaction design, product design, service design. For example, fashion is clearly human-centred, but so obviously so that it adds no value at all to parade or advocate its human-centredness.

The overuse of 'centred' is unfortunate because design isn't a shape and it can't have a centre. Design is a co-ordinated set of human activities, and far from being centred on anything, it is more like plate spinning. As design situations become more complicated, design teams have to keep more plates spinning. Choosing a centre and standing there is a guaranteed way to stand and watch most of the plates fall and break. Design doesn't need a centre. Instead it needs the agility to shift focus without losing the connections between ever shifting foci.

An ADS is abstract because it is distinguished solely by the types of choice that are made within it, and the connections between these choices. ADSs are intended to be a neutral 'post-centric' concept that can span all craft practices (e.g., graphic, furniture, moving image, textiles) and all design value systems. Thus as long as graphic designers and textile designers are committed to making the same types of choice, and the same types of connections between choices, then they are working within equivalent ADSs, even though their craft practices are distinct, as may be their choices of beneficiaries, evaluations or purpose. They are the same ADS when the types of design choices and interconnections are the same.
 
ADSs focus on choices within design. All design involves choices, whether explicit or tacit. Design Situations differ in the types of choices that are made. Simple design situations may only involve two types of choice. One type here will always relate to the features of the artefact being designed (e.g., consumer electronics, garment, public service, chair). The second type may relate to design principles, a design brief, a specification, or a designer's judgement. In each case, a different design situation results, because different types of choices are being made, e.g., choice of which principles to design to, choice of the scope and content of a brief of specification, or choice of which of their own feelings to respond to as designers.

More complex ADSs arise as further types of design choice are considered, e.g., choices of beneficiaries, choices of evaluation approaches and choices of design purpose. It is relatively straighforward to co-ordinate two types of design choice, but this becomes more challenging when three or more types of design choice are simultaneously under consideration. The ability to successfully and effectively connect between design options and choices becomes a major determinant of design success as design situations increase in complexity. The range of possible connections can become infinite (because we can connect to connections, as well as to choices), and thus only some connections can be given explicit attention within a design setting.  

The ADS concept provides partial support for assessing the coverage of design and evaluation approaches, which can support one or more types of design choices in one or more ways. At a very general level, the abstract work that design and evaluation approaches support corresponds to a set of meta-principles. These are 'meta-' in the sense that they are too abstract to guide design in specific contexts. Instead, they scope out the kinds of work supported by design and evaluation approaches, for example, generating options for any type of design choice, providing ways to sketch or otherwise communicate options, or establishing the credibility, feasbility or viability of specific options.
 
The advantage of a high level of abstraction is that it can support a wide range of analyses while using a relative small set of concepts. Currently, I have identified six meta-principles that scope out the kinds of work that frameworks of design and evaluation approaches need to support. The first is receptiveness, which relates to the ability of a design or evaluation approach to generate options. The second is expressivity, which relates to the ability of a design or evaluation approach to communicate options to project stakeholders. The third is credibility, which relates to the ability of a design or evaluation approach to identifiy valid options and choices. Within a specific Design Situation, these three meta-principles apply to every type of design choice that has been committed to, and every connection type that has been committed to (a connection type connects two or more types of design choices).

A fourth meta-principle of committedness relates to the types of design choice and connections that are explicitly committed to for a design setting. This meta-principle is the easiest to satisfy, since the choices required are very abstract. A design team may commit to make choices about features and purpose, but not about evaluations and beneficiaries. While this may appear negligent to self-proclaimed human-centred interaction designers, who feel obliged to make choices about beneficiaries and evaluations, product design companies such as Alessi and Apple do not depend on extensive market and user research (beneficiaries) or formal investigations on design quality (evaluations). The ADS concept makes it clear that not all design situations will involve the same types of choices. The meta-principle of committedness makes it clear that types of design choices are chosen. We'll ignore any resulting logical quandaries here, and observe that satisfaction of all other meta-principles is constrained out by committedness. Receptiveness, expressivity and credibility must all be focused on specific types of design choice and the connections between them.
 
ADSs thus result from strategic design choices arising from committedness. Two further meta-principles relate to the inclusiveness and improvability of a proposed design. The former is specifically focused on choices of beneficiaries and connections to these choices. The latter is dependent on choices of evaluations and the ability to understand and respond to evaluation outcomes.

This is all very complex, and very abstract too. Its value lies in the support it can provide (trust me!) for understanding, assessing and improving design settings on the basis of the six meta-principles and the types of design choice under consideration. For the latter, I currently consider only four types of design choices: artefacts, purpose, beneficiaries and evaluations. As a result, the scope and effectiveness of a design setting can be audited with recourse to only ten generic concepts (four choice types, six meta-principles). Similarly, frameworks of design and evaluation approaches can be audited in terms of the work that they support (e.g., expressivity, inclusiveness) for one or more types of design choices and/or the connections between them.

You can read a range of papers on my evolving ideas on meta-principles and Abstract Design Situations. However, many of these papers are experimental, exploratory, or both, and are thus often dense, abstract or both. The aim of this blog is to take time and make all of the above intelligible and concrete. This may take some time, so don't hold your breath.

I'm also hoping that this blog can become a resource for discussions on Design Situations and Meta-Principles for Designing that can refine, clarify and exemplify a high level framework for thinking about design practices and outcomes. In the meantime, you may find videos of presentations more accessible, two of my 2009 keynote at the Dutch HCI Conference, and one of my talk at Thinking Digital 2010 are available on vimeo. NB - there are two Part 2s for my Dutch keynote, the first one is Part 1! (the one with the slide of the Mac Airs).