Who is this book for

One of the most pressing needs today is to improve the quality of urban development for growing regions around the world — to maximize the benefits that urbanization can bring, while minimizing the potential problems and negative impacts for the future. To that end, this volume presents a so-called “pattern language” of a number of urban development best practices that have been identified through research at Sustasis Press, and with partners at KTH University, the University of Strathclyde, and a number of other universities, research centers, and individual collaborators.

This book also represents a contribution to a five-year collaboration with UN-Habitat on implementation of the “New Urban Agenda,” a framework document that seeks to maximize the human benefits of urban development at all scales over the period 2016-2036. The New Urban Agenda was adopted by consensus by all 193 countries of the United Nations in 2016, in an act of remarkable international unity. However, there remains an urgent need to implement its aspirations with effective, evidencebased tools and strategies. This volume is one contribution aimed at addressing that need.

This volume (and its on-line counterpart, npl.wiki) is also intended as only the start of a much wider international effort — not only applying the information herein, but seeking additions to it, and encouraging parallel efforts. Ultimately this and other projects may form a many-stranded network for sharing of the best available knowledge, tools and strategies

for better-quality urbanization. In that spirit, this work is by no means a “final word” — but it is our best curated formulation of the current state of evidence-based urban design and architecture today, expressed as a representative and (we very much hope) useful collection, for active builders, designers, planners, businesses, governments, and NGOs.

Why pattern languages?

In many fields today, pattern languages have been used successfully to develop and share best practice design tools and strategies. Perhaps the widest usage is in computer science, where pattern languages of programming (also called design patterns) are used to develop many operating systems, most games, and many other kinds of programs. In a remarkable spinoff, pattern languages also led to the development of wiki, which was created as a tool to share patterns of design, and later used (more famously) to create Wikipedia, as well as many other widely-used websites. Additional software spinoffs included Agile development, Extreme Programming, and Scrum methodology.*

Pattern languages have also been applied usefully in a surprisingly diverse number of other fields, including human-computer interaction, sociology, molecular biology, business management, manufacturing, and production engineering, to name a few. In fact, many thousands of patterns have been written, not only for software and computer architecture, but also for structural principles of organizations, education, social interaction, communication and information technology, even music, chess and poetry. Researchers in these disciplines have adopted the convenient pattern format to present their results, and were also encouraged to try and find links among their patterns. The pattern format is now embraced as a convenient standard in which to write new results in a variety of disciplines.

This diversity is all the more remarkable, considering that pattern languages had their origin in the built environment, and yet remarkably, the built environment remains one of the least well-developed fields for pattern languages. Therein lies a paradox — and an underdeveloped resource.

Pattern languages were introduced most famously in the 1977 book by the architect Christopher Alexander and his colleagues, A Pattern Language:

Towns, Buildings, Construction. The book offered three remarkable achievements all at once. First, it gave the pattern format for expressing a discovered design result in compact and logical form for future reference and distribution as an evolving best practice. The presentation occurred normally in seven parts: iconic name; representative (contextual) photo; links to previous patterns; problem-statement; discussion; conclusion (“therefore statement”); and links to subsequent patterns. Second was the idea of a pattern language in which the individual patterns link up using grammar-like rules. Importantly, this emphasized that design patterns are not

isolated entities, but are embedded in an essential web-network. Third, the book presented the specific collection of 253 numbered patterns developed by Christopher Alexander and his associates at that time.

What accounts for the usefulness of pattern languages across such a diversity of fields? They are in essence a way of capturing useful knowledge about the nature of a design problem, and expressing it in a way that can be easily shared and adapted to new contexts. However, the form of the knowledge is not rigid, but context-dependent and relational. This feature is especially useful for design problems that require very local and context-specific responses. Of course, this is very often the case for problems of urban design, architecture and building too. What accounts for the comparatively limited development of pattern languages in the built environment — the very field for which they were originally developed? One explanation is that some architects and urban designers do not like what they see as the book’s formulaic design guidance, which they believe constrains their creativity. That may be true for some, but by no means all. Another perhaps more relevant explanation is that, paradoxically, the very success of the 1977 book served to “freeze” the work in a seemingly immutable set of 253 patterns. The book became a best-seller, and an iconic work that some said must not be “tampered with.”

But that is not what the book itself said. On the contrary, the introduction made its aim very clear (emphases added):

Let us finally explain the status of this language, why we have called it “A Pattern Language,” with the emphasis on the word, “A,” and how we imagine this pattern language might be related to the countless thousands of other languages we hope that people will make for themselves, in the future… The fact is, that we have written this book as a first step in the society-wide process by which people will gradually become conscious of their own pattern languages, and work to improve them…

We hope, of course, that many of the people who read, and use this language, will try to improve these patterns — will put their energy to work, in this task of finding more true, more profound invariants — and we hope that gradually these more true patterns, which are slowly discovered, as time goes on, will enter a common language, which all of us can share...

You see then that the patterns are very much alive and evolving. In fact, if you like, each pattern may be looked upon as a hypothesis like one of the hypotheses of science. In this sense, each pattern represents our current best guess as to what arrangement of the physical environment will work... But of course, no matter what the asterisks say, the patterns are still hypotheses, all 253 of them — and are therefore all tentative, all free to evolve under the impact of new experience and observation. (Alexander et al., 1977, pages xv-xvii)

Unfortunately, with the exception of a few pockets of practice, this continued evolution has been woefully absent in built environment fields. Instead, the very success of the book has served to freeze its contents, protected even by copyright as well as by the practical difficulty of modifying or adding to printed pages. Yet the text above makes clear the intention to launch a much larger and more ambitious open-source project to develop many more patterns, and to edit, adapt, share, and apply them all.

* See for example Cunningham, W. and Mehaffy, M.W. 2014. “Wiki as Pattern Language.” In Proceedings of the 20th Conference on Pattern Languages of Programs (PLoP’13), Monticello, Illinois, USA (October 2013). 15 pages.

The proliferation of design patterns

This is precisely what has happened in computer science and other fields, with prodigious results. The comparatively weak results in environmental design fields are humbling — and instructive, for those who recognize the unmet potential of pattern languages to help to address new challenges for settlements.

For many years the kind of open-source exchange called for in the book was difficult to accomplish at any significant scale, since it required the cumbersome use of copier machines and the like. A few authors published compendia of new patterns, but without the ability to interact with and incorporate the original 253, they did not have a very large impact.

Of course, with the advent of the Internet, it became much more practical to share patterns, and even to turn the references that each pattern featured into “live” links that could be used to “click through” to other patterns. This is precisely what was done in 1987, not by environmental designers but by software engineers. In that year, Ward Cunningham created the “Portland Pattern Repository,” advancing both pattern languages of programming and their more famous outgrowth, wikis.

Both design patterns and wikis were developed to address a fundamental problem in software: simply specifying new solutions to new problems in sequence leads to a cluttering of code, and an increased likelihood of malfunctions from unforeseeable and unintended interactions. In 1987, Cunningham and his colleague Kent Beck, working at Tektronix Corporation near Portland, Oregon, were seeking new forms of software that would display what mathematicians sometimes refer to as “elegance”: the ability to do more with less. Cunningham embodied this principle in the question, “what is the simplest thing that could possibly work?” We enter process of exploration and adaptation, without assuming the need for detailed specifications in advance.

Cunningham was intrigued by the capacity of language, in its very ambiguity, versatility and economy, to serve more ably as a useful working model for problem-solving. A problem is, by definition, not pre-decomposed into simple functional units, but as Alexander noted, has many overlapping and ambiguous connections. Language mirrors this capacity, and therein lies its usefulness. Therefore, the goal is, in a sense, to achieve the same robustness of language, by endowing the model with its own set of powerful (but limited in number) generative components, much as language does.

Thus, the goal is not simply a matter of economy, but one of greater context-adaptive problem-solving power. In fact it goes back to the heart of Alexander’s concept of language-like networking: a simple grammatical system, functioning generatively, can be far more powerful than a complex set of specification-based processes. As Cunningham put it, when asked by programmer Tom Munnecke to explain how “the generativity of a pattern is a way of expressing complexity:”

That was an idea that excited me, and that seemed more powerful than most... And that is, language is generative. And that idea that I can have a set of rules that generates something that I could value is really important. So the question was, why don’t we do everything that way? And the answer was, well we pretty much did, until we let professionals get involved. …And they made it complex by trying to make it simpler, because they didn’t understand how some system of rules could generate behaviors instead of specifying behaviors.

This generation refers to the capacity to reproduce the essence of a functioning structure, without having to specify all of its characteristics. A simple example is the distinction between the way a genetic process generates the blue eyes, say, of a child, which recapitulates the blue eyes of the parent without having to specify them in minute detail — their intricate retinal flecking pattern, precise round shape, and so on. Instead, the genetic process is able to generate, and regenerate, an intricately complex structure from a relatively simple set of language-like (or recipe-like) instructions.

So it is with pattern languages and their patterns. The goal is to pick out the most salient features that are needed for regeneration within a specific context, and to establish a generative process that uses those patterns. This process is very much like the way older cities and buildings were traditionally generated using linguistic concepts, often without the need to state them explicitly. In the case of pattern languages, the process is only formalized, so that designers can be more articulate about the needed design aspects, and so that the result can be more successful, more durable, and more sustainable, responding to the best available evidence, and representing a best adaptation to human need.

The continuing need for pattern languages in many fields — including the built environment

Just now, by contrast, the human species is drowning in overly complicated and malfunctioning designs, from a human point of view. They may be exciting, they may be stimulating, they may be entertaining — indeed, they may not be malfunctioning in the short term, but instead, offer great power and allure. But we are like the fabled Sorcerer’s Apprentice, unleashing a power we cannot control. Especially in comparison with the durable structures of nature, and of our own history, the results lack longterm resilience and sustainability. We can enjoy them, we can marvel at them, we can admire them — but we must also commit ourselves to deep reforms.

An apparent paradox is that today we are able to produce more volume of building than ever before in human history. Indeed, we are in an era of unprecedented urbanization, on course to build more sheer area of urbanization in the next fifty years than in all of human history. It is therefore a matter of highest urgency to address the nature of this urbanization, and its long-term impacts on economy, technology and quality of life — and to determine the levels at which reforms are needed in policy and practice.

It is a thesis of this book that those levels are very deep indeed. At the heart of the pattern language methodology is a recognition of changes needed in the very nature of technological methodologies, and the inadequate feedback capacities of our current linear systems — particularly as they impact the use and depletion of resources, the systems we use for developing and applying adaptive knowledge, and related challenges.

This is also the reform-minded insight behind the related movements of Agile, Scrum, wiki, and other innovative reforms in methodology of design, and technology more broadly. Pattern languages, as we have discussed, played a role in shaping these other movements. The stunning, if partial, success to date hints at more to come, and suggests that the full potential of pattern languages — especially in the built environment — has not been reached.

One of the great advantages of pattern languages is that they do contain within them the capacity to establish reciprocal feedback channels through their web-networks of hyperlinks. The implications of this capacity are broad, although a full discussion is beyond the scope of this book.

Why these patterns in particular?

This volume is not meant to provide an exhaustive library of patterns, but rather, to provide a representative curated compendium of relevant new patterns, suggesting the potential for many more. The book is divided into three sections representing places, networks and processes (“patterns of scale,” “patterns of multiple scale,” and “patterns of process”) with a series of pattern groups under each, and four representative patterns in each group. The selection of four is not significant, except as a means of including a small but illustrative sample of each kind of pattern.

There is also an on-line companion “repository” that includes these patterns as, in effect “seed patterns,” which can be edited, deleted, added to, and used in any other way desired. It can be found at npl.wiki. We hope very much that this on-line version will lead to the evolution and use of many more patterns.

The patterns curated herein are not the only ones that are possible, certainly — and indeed, many regions are using very different patterns today (for better or worse). But a key purpose of this book is to show with the patterns herein examples of a more reliable, evidence-based approach to sustainable, resource-efficient urban development, promoting a higher quality of life, and at the same time, a healthier and more sustainable form of economic development. We document this claim with numerous research citations within the patterns, and we further demonstrate this claim by showing some concluding examples of several contemporary

cities that do incorporate these patterns very successfully, with measurable economic, social and ecological benefits.

Other users may assert other patterns, or dispute the patterns we propose — and that is fine. All that is required is that the preponderance of evidence over time shows which patterns succeed best on local human terms, including social, economic and environmental dimensions. We want to know not only which patterns seem to be universally more beneficial, but also locally more beneficial in many different contexts, and expressing many different adaptations and variations — and in some cases, wholly new patterns, that may prove useful elsewhere. In this way, the entire collection of patterns can grow more useful, and at the same time, more diverse and extensive. This is, after all, how science works, and how knowledge works.

Ultimately, we may have many different collections of patterns, some sharing common patterns, some slightly different, and some altogether different, based on context. This is the core idea behind “federation,” the concept that motivates the new “federated” generation of wiki, of which the new wiki is an example. This would deliver on the promise of the original book, to see “countless thousands of other languages we hope that people will make for themselves, in the future.”

Why this particular format?

For the printed version, we have chosen to stay relatively close to the structure and appearance of the original pattern format introduced in 1977, for three related reasons. First, other versions of pattern language structure and appearance have been developed for the built environment, in part following the invitation of the book — yet no format has proven as user-friendly, as appealing, or as effective, as the original. This project is nothing if not evidence-based, and reliant on what has been shown to work — unless and until other, better practices are demonstrated to be more successful. Second, a consistent form of pattern is necessary so that the original goal of shared languages may be possible, and so that the patterns may be re-used in various customized project languages, working together. Third, the original book itself provides explicit direction (emphasis added):

…we have written this book as a first step… We have spent years trying to formulate this language, in the hope that when a person uses it, he will be so impressed by its power, and so joyful in its use, that he will understand again, what it means to have a living language of this kind. (A Pattern Language, pp. xvi-xvii)

Echoing our second point, the book also makes clear that the format presented is intended as an essential characteristic of the project:

There are two essential purposes behind this format. First, to present each pattern connected to other patterns, so that you grasp the collection…as a whole, as a language, within which you can create an infinite variety of combinations. Second, to present the problem and solution of each pattern in such a way that you can judge it for yourself, and modify it, without losing the essence that is central to it. (A Pattern Language, p. xi)

The success of the format is clear, and the invitation to use and modify is clear. This book is therefore one step in response, and an invitation to all those who are serious about further open-source development of pattern languages and related advances. Let us join with many others in a larger collaboration, now for the built environment as well — precisely as called for in the original book.

Patterns for a New Urban Agenda

Many of the patterns collected here are also suggested by the New Urban Agenda, the aforementioned 2016 international framework agreement on urbanization adopted by acclamation by all 193 countries of the United Nations. The document places heavy emphasis on the role of public spaces, including streets and sidewalks, as essential elements for healthy urbanization. It articulates this new priority for public spaces in no fewer than ten paragraphs.

For example, Article 37 promotes

... safe, inclusive, accessible, green and quality public spaces, including streets, sidewalks and cycling lanes, squares, waterfront areas, gardens and parks, that are multifunctional areas for social interaction and inclusion, human health and well-being, economic exchange and cultural expression and dialogue among a wide diversity of people and cultures…

The New Urban Agenda also emphasizes the economic importance of public spaces, as in Article 53:

We commit ourselves to promoting safe, inclusive, accessible, green and quality public spaces as drivers of social and economic development, in order to sustainably leverage their potential to generate increased social and economic value, including property value, and to facilitate business and public and private investments and livelihood opportunities for all.

The New Urban Agenda also emphasizes the interconnected, “network” character of public spaces, with special emphasis on streets as public spaces, and the ways they and other public spaces connect to private edges. For example, Article 100 supports

...the provision of well-designed networks of safe, accessible, green and quality streets and other public spaces that are accessible to all and free from crime and violence, including sexual harassment and gender-based violence, considering the human scale, and measures that allow for the best possible commercial use of street-level floors, fostering both formal and informal local markets and commerce, as well as not-for-profit community initiatives, bringing people into public spaces and promoting walkability and cycling with the goal of improving health and well-being.

Finally, a number of other articles in the New Urban Agenda emphasize the integration of public spaces with other key characteristics of urban form, including “polycentrism” (many regional centers with a full mix of housing, employment and recreation). For example, Article 51 supports

...the development of urban spatial frameworks, including urban planning and design instruments that support... appropriate compactness and density, polycentrism and mixed uses, through infill or planned urban extension strategies, as applicable, to trigger economies of scale and agglomeration, strengthen food system planning and enhance resource efficiency, urban resilience and environmental sustainability.

An evolving theory of urban form, based on an evolving science of cities

This emphasis on public space frameworks organized around streets and their active edges, and around mixed use, polycentrism and compactness, reflects a notable shift from the dominant 20th Century (mostly European and American) urban theories. These older theories, rooted in an earlier industrial model of cities, have given way to a more dynamic, more complex view of cities — one that also reflects new scientific insights from the biological sciences, and from other advancements in mathematics and other fields.*

The form of many cities around the world today is still dominated by these older models. It must be recognized that these models have proven effective in supporting rapid urbanization and economic growth, and in removing millions from poverty. That achievement should not be minimized. At the same time, the older models rely on unsustainably high rates of resource consumption and depletion, and related long-term consequences like pollution, greenhouse gas emissions, climate change and other potentially disastrous long-term impacts. We can also see that, for all its gains, the 20th century paradigm has also been socioeconomically costly, by segregating and essentially trapping many of the poor. The evidence increasingly points to the need for a major transition to more resource-efficient forms of urbanization, and of technology — and to urbanization that also more efficiently delivers better long-term quality of life for human beings, without the many negative impacts of the older models.

Accordingly, the patterns herein reflect this new view of cities, and indeed the new understanding of the inter-connected, web-like patterns within cities — a view on which the technology of patterns itself is based. Therefore, in accordance with the New Urban Agenda, the patterns here describe compact, polycentric urban development, public space frameworks, a mix of uses, multi-modal forms of transportation over well-connected, walkable street systems, active street-level building edges, human-scale design, ample greenery and natural characteristics, and other related specifications. Again, these patterns are not the final word, but they do reflect our best current formulation of the state of the urban science, and the lessons for urban best practice.

* These topics are discussed in much greater detail in Mehaffy, M.W. and Salingaros, N.A. (2014), Design for a Living Planet, Portland: Sustasis Press.