At its broadest level this thesis is concerned with the design and planning of cities. This chapter draws together the two main strands of the thesis, the nature of urban structure and the design debate. The last four chapters have been mainly concerned with the description and analysis of urban structure. This chapter now switches to the issue of design – which may be regarded as the other side of the same coin - and builds on the understanding of the nature of urban structure drawn from the preceding chapters.
Already the thesis has suggested interpretations of connective networks and connector streets, and of the complex, characteristic nature of traditional street patterns. These represent forms of urban structure now considered desirable by urban designers and planners. What is now required is a means of achieving these through design. In the last chapter, we have seen how the way in which structures are formed affects their resultant shapes; how 'programs' influence 'patterns'. This chapter therefore considers the nature of different processes of design, or 'modes of creation'.
It then goes on to investigate the principles of a constitutional approach - design by program rather than by pattern - in the context of contemporary transport provision. The chapter finally concludes by drawing all the design components of the thesis together into an Integrated Design Framework.
10.1 Modes of urban creation
There is a wide variety of ways in which cities may be designed, or in which city design may be interpreted. Kevin Lynch describes three broad normative models of city form, the cosmic, the machine and the organic.
With the cosmic model, the city plan uses urban structure as an expressive medium, somewhat akin to the architectural expression of the form of a building. This expressive model may be equated more generally with the 'city as a diagram' or the city as a work of art .
The city-as-machine model is a more functional construct. Yet this might also be equated more widely with the idea of city structure as a direct expression of technology, where the form is significantly influenced by the characteristics of the structural elements themselves. Part of the interpretation of city structures reflects this aspect, though part also reflects 'artistic' expression of this technology.
The organic model may be interpreted in many ways, and has a resonance with the complex, interconnected, 'organic' forms of traditional urban design which often provide exemplars for contemporary neo-traditional urbanism. For this reason, we shall concentrate for now on the organic model, though this exploration will ultimately be related back to the other two.
10.1.1 Organic Models and Modes of Creation
The prevalence and recurrence of the organic metaphor in planning and design literature is quite remarkable - not least because of the eclectic roll-call of those proffering an organic approach. In addition to the 'Organic Family Tree' of Howard, Geddes, Mumford, Jacobs and Alexander we can add classic modernists like Lloyd Wright and even Le Corbusier.
Organic interpretations range from the most general association of urban systems with living or evolving entities through definite associations of settlements with living organisms to specific analogies of form, where urban components are equated with living tissue, organs, cells and nuclei. Regarding urban structure, we have already seen a variety of skeletal references, and to this we can add the classic mixed metaphor of the main traffic 'artery' being the 'backbone' of a settlement.
The organic metaphor has been dissected by a number of authors. Here, some key facets of direct relevance to the design of urban structure shall be focused on.
We have already seen organic patterns are nowadays considered worthy of emulating, though previously they were seen as part of the problem. Nowadays we see attempts to replicate traditional patterns, and to recreate what might be described as 'organic' urban forms.
However, these neo-traditional exemplars are not necessarily created in an 'organic'
manner: 47339d89f32c310ff4147addb1127eb1they may attempt to create organic patterns, but not necessarily in an organic way.
We can see this, for example, in the design of Poundbury , which was inspired by neo-traditionalist and indeed organic ideas.
The master plan may show the Poundbury development forming a kind of pseudo organic pattern. However, on inspection, this appears to be simply imposed on the existing site, oblivious of the existing pattern of fields and farms. Indeed, the plan exposes a striking imposition: the Poundbury development, as ultimately envisaged, would obliterate an existing Roman Road that runs through the site.
Where once there was a dead-straight, connective, historic route running out from the town, there would now be a 'village' of houses arranged around a Street pattern of contrived irregularity. Despite its architectural success in integrating buildings and highway forms internally, Poundbury therefore cannot be seen as an 'organic' extension of the town of Dorchester. It is a design conceived elsewhere, and parachuted down, as it were, from the drawing board.
On one level, the case of Poundbury illustrates that organic patterns are not necessarily the same as those created by organic processes. This is consistent with the comparisons noted earlier between traditional and neo-traditional layouts.
At another level, the question of process opens up another perspective on organic creation: if a city is an organism, how was it created in the first place? Who designed the' organism'?
10.1.2 The genetic paradigm
Lewis Mumford, contemplating a series of medieval town plans, noted that although each was unique, according to its unique situation, the plans seemed together to present a recognisable pattern. We might say that these displayed 'homologies of form' which Mumford attributed to their functionality. Mumford reflects:
"The consensus is so complete as to the purposes of town life that the variations in detail only confirm the pattern. The consensus makes it look, when one view a hundred medieval patterns in succession, as if there were in fact a conscious theory that guided this town planning".
Mumford's comments invite the thought that the towns' functionality made them
appear 'too well planned' to have arisen by chance. This sentiment has a resonance with the 'argument from design' in biology/theology. This is the view that the intricate harmony of the workings of nature must be explained by the existence of a conscious designer.
However, as Steadman has remarked:
"Just as Darwin inverted the argument from design, and 'stole away' God as designer, to replace Him with natural selection, so the Darwinian analogy in technical evolution removes the human designer and replaces him with the 'selective forces' in the 'functional environment of the designed object".
Mitchell echoes this point:
"In other words, you can get acceptable results by combining smart designers with dumb critics, or by teaming smart critics with dumb but energetic designers. You may prefer God(a smart designer with no need for a critic) or evolution (indiscriminate generation but deadly effective criticism)".
A problem with the conventional organic model of city design is that it tries to say that the city is a living organism, but this is an organism preconceived on the drawing board. If we view the problem from this point of view, we effectively change from an assumption of design - the city as a designed object - to the more ambivalent territory of whether and to what extent a city is designed at all. In fact, we can shift to a 'genetic paradigm', where we effectively have a polarization between an 'evolutionary' explanation and a 'creationist' one.
In the perspective, species evolve over time, typically from simpler to more complex forms, without overall design or direction. This has a parallel with 'natural', 'unplanned' cities, which arose in a bottom-up fashion - architecture without architects.
In the CREATIONIST perspective, however, a creature is conceived as a purposive act, its form settled in the beginning. In this perspective, the city is a designed object, like a building, with a preconceived 'finished' form.
Conventional town planning is 'creationist'. It invented the idea for the need for a town designer or planner. It created planned towns, typically in a top-down manner.
Alexander evoke the creationist spirit where "...each act of design or construction is an isolated event which creates an isolated building - 'perfect' at the time of its construction, and then abandoned by its builders and designers forever". And Greed alludes to the idea of the bad old days when urban design was "very much a top down process in which the 'designer' looked down on 'his' drawing board, taking the 'God' view', and thus created the Grand Design".
Frederick, master planner of Harlow New Town - who considered that the 'town is a living organism' - describes his own Master Plan for Harlow:"the three lines of communication - road, rail and river - form the base-line of the new town, from the centre of which (the site of the railway station) a semi-circle is struck to give the most compact theoretical perimeter for the town".
This sounds like nothing other than the planner-God, at the drawing-board, in the act of creation.
In the case of a conventional Master Plan for a 'planned' town, even where the
intention is to allow 'flexible growth', it is the designer who has consciously decided the relationships between the parts and the whole, what will be the centre, and what will be the subsidiary parts:
In terms of circulation, it shows a network of principal roads designed as a hierarchy... it shows the routes of public transport, the relationship of bus and railway stations to the town centre and railway sidings to industry; it will define the principal pedestrian routes about the town and their connection with traffic nodal points, such as a bus interchange station. It does not attempt to give the layouts inside the built-up areas.. .".
This type of master plan predetermines the major relationships, providing a strategic framework - or straitjacket - within which local design must take place.
Of course, not all planned cities are unique creatures generated of their sites; they may be adaptations of existing theoretical models. This gives rise to an intermediate case between the pure creationist and pure evolutionary modes of creation. Again, biology gives a suitable analogy: this time with a pre-Darwinian interpretation of evolution.
The ARCHETYPAL perspective sees the city as a transformation of a basic archetype, which embodies a set of features possessed by all 'species'. This may then be flexibly applied or moulded to the site, as with a model 'linear city' or a cluster of 'garden cities'.
However, the configuration of relationships between parts, with respect to the whole, is prescribed, and is expected to remain the same over time. Thus, relationships between the central city and the satellite city, or the relationship between the main spine of a linear city and the transverse side roads, or the 'beads' and the 'string' are fixed in advance. The final form, in a sense, is the starting point. In this sense, the archetypal mode, though apparently flexible, is still a top-down imposition: the city is still a tree.
This diagram depicts a contemporary recommendation for the structuring of cities. The graphical resonance may be coincidental, but the underlying implication is not. In both cases the structure is devised from the outside in, or from the top down; the parts are organized in a specialized and subordinate relationship to the whole. And, as with Howard's Social City, the fact that the layout is diagrammatic, allowing local compositional license, does not alter its underlying logic - that of Alexander's Tree.
The three modes of creation are presented in Figure 10.3 and Table 10.1. Here, the archetypal mode is placed as an intermediate type. However, the 'archetypal' interpretation does not really explain the origin of the archetype, except in creationist terms. In fact, within the genetic paradigm, the city-as-work-of-art, city-as-machine and city-as-organism become variants of the same thing. It is in this sense that the city-designed-as-an-organism is not organic. Indeed, in treating the city as an organism, planners were treating it as if it were an organism of machine-like simplicity. Although the rhetoric may have been of cells, the reality seems to have been crude 'organs' with the simplicity and specialism of machine parts (e.g., heart as 'pump', etc). This means that, for all its anthropomorphic contrivance, La Ville is more like a machine, after all, than an organism.
Whether the geometry of the town plan is considered as an artistic 'diagram' or a
mechanistic 'blueprint', or whether the product is supposed to be a living thing, the process seems to be the same: that of 'creationism'. We need to look elsewhere for a more organic, evolutionary approach.
Central city and the satellite city, or the relationship between the main spine of a linear city and the transverse side roads, or the 'beads' and the 'string' are fixed in advance. The final form, in a sense, is the starting point. In this sense, the archetypal mode, though apparently flexible, is still a top-down imposition: the city is still a tree.
10.1.3 The Quest for an Organic Approach
If we desire to create neo-traditional urbanism, or organic urbanism, and avoid the pitfalls of 'monolithic modernism', what would be the appropriate approach? It seems that contemporary urban planners and designers are aiming for some kind of traditional values and virtues (eg, streets, mixed use, inter connective patterns), sometimes backed with aspirations of organic overtones of wholeness or coherence.
The characteristic structure of traditional f8wNkBmkQ2Pf+v3rqQVWYGQ==orms seems to be significantly explainable in terms of the ways in which they grew. Traditional urbanism often, though not exclusively, can be seen to be based on an evolutionary mode; urban structure would grow with no preconceived overall pattern; hierarchy would be emergent, not imposed. There was no conscious imposition of ('formless') patterns.
However, what is required is not necessarily anarchy or a complete absence of rules. The marvel associated with Rudovsky's traditional organic settlements is surely not It has been stressed that Howard's Social City was a diagram, not a plan. This is consistent with Social City being interpreted not as 'creationist' but 'archetypal'.
They are chaotic. Rather, it is surely because they have some sort of coherence, and yet this coherence was achieved without master planning. The coherence may be explained by the use of some sort of bottom-up approach, building up settlements incrementally, within a given design tradition.
We have already seen that conventionally recognized features such as 'centers' and 'radials' may be interpreted as emergent effects, not prescribed by any global rules, nor explicitly present in the local rules. Similarly, districts can be regarded as being emergent. Designing cities using these as a starting point does not seem compatible with an evolutionary approach. Reverse-engineering the city using Lynch's elements - nodes, paths, districts, edges and landmarks - as top-down, compositional expressions would seem to end up with a creationist approach like any other.
What seems to be required is a bottom-up process - but not one that uses districts or neighborhoods as building blocks. It seems that more fundamental units are required, at the level of streets, buildings and plots. We have already seen how the simple addition of routes can build up quite complex patterns, that give rise to recognizable 'wholes', despite not being co-ordinate at the global level. The patterns in Figure 9.11, while showing some degree of variation, seem to have a family resemblance - this is because they were generated by the same 'program'. The presence of a program suggests some degree of consistency at an appropriate level.
This suggests that a program-based rather than pattern-based approach is appropriate. Indeed, the New Theory's self-proclaimed most controversial feature is 'to generate urban structure without a plan' such that 'in some fashion, the large-scale order will emerge, organically, from the co-operation of the individual acts of construction' (Alexander et a!, 1987:37). This fits with the evolutionary mode of creation. With this approach, there may be an underlying program, but no fmal pattern is preconceived. Indeed, in the design exercise illustrating the New Theory, the street pattern unfolds in an incremental way, giving rise to a textbook example of characteristic structure.
A similar approach may be found in Sorkin's Local Code, in which local rules may be used incrementally to build a city. The code recognizes that a vision already concretized pre-empts the greater possibilities of an incitement to open many interpretations.., it seeks a city designed not simply through the deductions of a dominating generality but also via induction from numberless individual points of departure.
Local Code is written like an abstract manifesto - there is deliberately no illustration suggesting any design outcome. Rather, it comprises a systematic set of rules covering a vast range of city-building concerns; patterns of routes are built up using explicitly defined branching algorithms. This is a very pure case of a program.
Sorkin describes his code as a 'kind of utopia'. However, such a code need not be some remote or fanciful ideal. A program-based approach is right here with us. Indeed, it is the approach of the highway engineer.
10.1.4 The local code of the highway engineer
The caricature of the 'romantic' urban designer or planner dreaming up artistic patterns for their town plans is familiar enough. Unwin refers to town planning as an art and the town plan as a canvas, considering the urban designer as an 'artist', with a refined 'palette' of street types. Gibberd similarly refers to the art of town design and alludes to painting and sketching. We can imagine the designer at the drawing board, poised with palette, as a bit of this grain is mixed with a bit of that, some texture added here and some theme elaborated there.
By contrast, the engineer is stereotypically the technician, devising standardized layouts from the engineer's manual, as urban structure is fitted together with so many joints, channels, and connections: the allusions here are to construction parts, plumbing or electrical circuitry. If anything, surely the engineer's city is a machine.
The above caricature seems to have some ring of truth to it. After all, it is planners who dream up 'visions', while engineers do the calculations to make these concrete. In terms of urban structure, there is the perceived problem that the highway engineer has control of form and pattern, and the implicit criticism that unimaginative technicians are stifling the creative potential of more artistic designers.
However, when looking more closely at the generation of urban structure, things are not what they might appear at first sight. The highway engineers' rules relate only to local connections, and therefore allow an incremental approach where no single configurational outcome is prescribed. The incremental approach also allows flexibility of application by individual designers and allows emergent forms.
By contrast, predetermined planning visions, conceived of as a whole, may be equated with top-down imposition. In this sense it is more likely to be the rigidity of the master planner's vision which constrains the creativity of the site planner or architect 06d4e8e2e58679268a48f88556662512when it comes to detailed forms, rather than the highway engineer.
The rational, rule-based approach of the highway engineer could yet precipitate an organic renaissance, allowing the creation of natural, emergent forms, while the master planner is implicated with rigid authoritarianism.
With the boot on the other foot, as it were, one could point to the crudeness and oversimplification of many planning visions. These conjure up not so much the image of the artist's canvas, as planning being a glorified version of 'painting by numbers', where old-style zoning conventionally indicated by standard color washes has rigid boundaries and little scope for personal interpretation.
It is therefore possible to synthesize street pattern using a series of local rules, without specifying any overall pattern. This is a constitutional, rather than a configurational approach. Seen this way, the apparent lack of design guidance for overall pattern, encountered in Chapter 4, is perhaps better interpreted simply as an absence.
As for networks, so it goes for route types. Buchanan explicitly rejected the notion of road types such as 'ring roads' which presupposed the final layout of a road with respect to overall network structure:
Instead, he opted for route types defined only by their local roles and connections. In this sense, Buchanan's was an adaptable 'evolutionary' mode of creation, rather than a more rigid 'archetypal' one. It is tempting to suggest that this is part of the reason why Buchanan's system has been so resilient, and its descendants live on today, albeit in an adapted form, almost four decades later (while Traffic in Towns' physical planning and urban design solutions have fallen out of favor).
These considerations suggest that the strengths of the current hierarchy should be built on, not undermined. Te3ac317776d932c543a30ab7e93fdc20his means the flexible - constitutional - parts, albeit applied in a consistent manner (i.e., program). This means that there is no need for top-down imposition of pattern (Alexander's Tree), standardized or otherwise, but there should also be no spurious fixing of constitution to tree-like configuration (Buchanan's Tree). In fact, 60 Buchanan's proposed route types were almost purely constitutional - the route types could almost be defined independently of function, but purely in the sense of which route type was allowed to connect to which other type. Buchanan's types avoided both the compositional overtones of 'ring roads' and 'tangential roads' and the configurational overtones of 'spine roads'. However, awareness of the configurational role of routes (as discussed in this thesis) need not imply preconceiving the configuration of the network. An iterative process relating route type, structural role and functional significance will take place as the routes 'work themselves out'.
In contrast to the ambiguity of Buchanan's Tree, the precisely constitutional part of Buchanan's approach can be seen as the antit3d123f64e9c9574d91cf9dfccf37ee9fhesis of, and perhaps the antidote to, Alexander's Tree.
However, advocacy for this does not imply descent without some modification. Firstly, there is a need to promote greater connectivity in networks, to make these more amenable to pedestrians. This could be done by a few extra rules or relaxations. Modifying just one or two rules of allowable connection or junction type could open up a wide variety of new permutations of overall structure.
Secondly, the problem with Buchanan's program was in its modernist tendency to promote a disassociation of roads and built form, a system ideal for route propagation and point blocks. Therefore, although Buchanan provides a robust and adaptable morphogenetic program, its denial of the street still marks it as a disurban mode of creation.
Hence there is a need for the program to have the land use and urban function 'built in' to it, rather than being appended to it. Together, the implication is that by a system of local rules and incremental design it might be possible to (re)create functional urban development, and that even though simple rules are used, organized complexity may result. The 'complexity' part allows for the kind of variety, interest and stimulation desired by planners and urban designers, while being sufficiently 'organized' - as an aggregation of simple workable rules - that it itself is workable. This would mean that there would be sufficient 'structural logic' to satisfy the transport planner and highway engineer (e.g., limiting access to major traffic arteries) while also implying the kind of hierarchy that is acceptable to the planner (e.g., allowing diversity and contrast between different types of route and space).
Hence, by design of component parts, i.e., routes with their appended frontage functions, it is possible to build up street pattern, and hence urban structure, in an organic manner. The implication is that a constitutional, program-based approach to design may be not only sufficient but desirabliPX3s0tNLRyI5pNu7r76REeFoWdyDDA45sLWV3goeW4=e.
10.2 Constitutional Structure
Having considered the general form of cities, we now return to the detail of their structural connections, and a constitutional approach to their design.
A constitutional approach implies two essential properties: firstly, rules that relate to local relationships, and secondly, the area-wide application of those rules. The first property ensures there is variety and no constraint on the ultimate configuration; the second property maintains some sort of consistency. Taken together the result should be something between total conformity and total chaos.
Without the second property, we would effectively have no consistent approach at all - and lack even the loose cohesion of 'unplanned' settlements, where local tradition acted as the area-wide influence. Here, the scale of the 'area-wide' approach is left open: this could imply national, regional or district guidance. This nested approach is quite compatible with present conventions in both highway engineering and town planning.
We have seen that the conventional highway engineering approach, such as that derived from the Buchanan report, may be part of the solution, in that relationships between route types are specified, but no overall configuration is presupposed.
What shall be proposed here is to build on this existing constitutional approach, but adapt it to contemporary needs, particularly to accommodate sustainable modes and urban functions more explicitly. The post-Buchanan layouts have struggled to recreate 'vital' urbanity. The proposal here is to evolve a more complex constitution which would be nearer to replicating both the processes and products of traditionally evolved urbanity. The result may still be 'artificial', but much closer to the level of artificiality that is possessed, more or less, in any 'natural' town.
10.2.1 Hierarchy and connectivity revisited
The basis of the constitutional approach here is the local rule system for relationships between route types. This section brings together the issue of hierarchy and the potential for local code as outlined.
The hierarchical matrix of Chapter 5 presented a 'periodic table' of route types, related to each other according to modal or frontage function, but without finally expressing the allowable connections between those types. Clearly, we could retain the convention by which routes may only connect to those adjacent to each other in the hierarchy - but this would seem to defeat the point of having a flexible matrix of types. On the other hand, we could have a completely laissez-faire situation where any type was allowed to connect to any other - although this might defeat the point of having a discrete typology in the first place. To explore other possibilities, we need to investigate hierarchical connectivity in more detail.
Three ways of expressing the conventional hierarchical constitution, elaborated in terms connections between vehicular route types. Stated verbally, An access road may connect (only) with another access road or with a local distributor; and a local distributor may connect (only) with an access road, another local distributor or a district distributor; and so on.
However, the diagram also embodies a second kind of connective rule, one not necessarily expressed explicitly in design guidance, but a feature of conventional road systems nonetheless: this is the condition of 'arterially'. This says that a local distributor must, at least, connect with another local distributor or a district distributor: connection to access roads is neither necessary nor sufficient. At first sight this may seem to be a trivial detail, so trivial as to not even appear in the 'small print' of the regulations. However, it turns out to be significant, as it helps unpack the notion of 'hierarchical' connectivity, into two distinct components, relating to a kind of accessibility and arterially.
Accessibility here relates to the allowable connections between route types. Conventional road hierarchy imposes access constraint, by forbidding certain connections, e.g., an access road may not join a primary distributor. This is a familiar enough constraint, and can lead to a local loss of accessibility where connections are severed.
Arterially relates to the necessary connections, or contiguity between route types. Arterially ensures that each road is connected to (contiguous with) the network of roads at its own level and above. As we saw arterially does not of itself restrict accessibility, since an unclassified road may join an 'A' road.
Arterially sounds a plausible enough principle, and restraint on accessibility has some justification. However, what conventional hierarchy does is to combine both access constraint and arterially, applying it to the urban road network. We may accept this for the moment, but what are the alternatives?
10.2.2 Constitutional types
From the combination of the two properties accessibility and arterially, we can generate four permutations of constitutional type.
The MOSAIC case implies a 'non structural' hierarchy. This could mean 'no hierarchy', but it would also include cases where routes are differentiated, but are not consistently ordered in any recognizable pattern of connectivity. This could represent an urban designer's 'spatial' hierarchy. The DIFFERENTIAL case may be found in traditional settlements whereromk4/fSyYX8yTwB2bJdJg== major streets have joined up to form through routes, forming a 'natural hierarchy'. The TRIBUTARY (or 'hierarchical') case is effectively that used in conventional road hierarchy. The LABYRINTHINE case is less immediately recognizable as any real system, but could represent cases where a 'hierarchical' system was incompletely implemented.
Table 10.3 generates an additional typological set to previously recognized types based on composition or configuration. Since constitution is conceptually independent of configuration, we can see that their association, in Buchanan's Tree, is not inevitable.
Note that the constitutional definition of tributary neatly encapsulates the sense of the word as previously encountered, even though tributary layouts may at first sight appear to be associated with configuration or composition (e.g., 'tree-like'). Yet, the 'tributary' label in Chapter 8 alludes to relative depth, which may be built up by branching or circuitous layering, and can therefore be regarded as being independent of layout.
We can also see how it would be possible to build a design strategy based on arterially and accessibility, where the rules would reside in the allowable connections between route types - without having to refer to overall pattern at all. This would form the basis of a constitution-led approach, where local rules would lead to emergent complexity in configuration. This would allow Alexander's Tree to be overcome.
The question then becomes one of how to build on the conventional case: how may accessibility and arterially be related to contemporary needs, in particular, different modes?
10.2.3 Principles for modal fit
This section suggests judicious combinations of arterially and accessibility for the three 'key' modes. We can accept for the time being that the conventional system (tributary) is appropriate for the car and general traffic. What, then, for the pedestrian and public transport?
For the pedestrian, we would not wish to constrain accessibility, in the interests of maximizing connectivity: therefore, any minor pedestrian route should be allowed to connect to any major pedestrian route. Access constraint also seems counter-productive for the public transport system, where this would imply a layered or branching system requiring several changes to reach the highest level services. However, arterially would be positively beneficial for a public transport system, since it ensures that any service connects with the wider network, ultimately upwards to the national level. For the pedestrian network itself, arterially seems unnecessary, since a pattern of pedestrian routes of varied intensity or importance, juxtaposing busy thoroughfares and quiet lanes, seems quite natural.
This suggests that: the MOSAIC case is appropriate for the pedestrian, and indeed, the layout shown in Table 10.4(a) could be interpreted as a walk-based historic layout;
? the DIFFERENTIAL case is appropriate for public transport and the pedestrian, and indeed, the layout shown in Table 10.4 (b) could be interpreted as a traditional settlement in which the main streets become through routes served by public transport;
? the TRIBUTARY case is the familiar car-oriented layout, as the layout shown in Table 10.4 (c) suggests a traditional town converted to a 'hierarchical' system.
These distinctions imply that a superposition of constitutional types is appropriate, rather than lumping public transport and the pedestrian together with general traffic. In the conventional road hierarchy, while the pedestrian and public transport may not have explicit constitutional relations spelt out, it can be seen how the tributary system would be disbeneficial. In fact, when we look at existing reality, we can see some potentially adverse tendencies.
While it is argued that arterially is not necessary for pedestrian networks, some planned pedestrian systems may have a kind of arterially implied. Where we have a completely separate pedestrian system, it is as if this mirrors the road system, where it is the principal pedestrian routes - including 'pedestrian spines' - that are consciously configured, into a contiguous network that possesses arterially. Where the pedestrian system occupies the lowest rung of the vehicular hierarchy, it also possesses a kind of arterially as far as this goes (i.e., the access road network), albeit that the access constraint then kicks in to deny a wider connectivity.
For public transport, there has traditionally been a lack of co-ordination or formal hierarchy: different modes, operators and infrastructures have seen to that. The resultant lack of arterially would be exemplified by the case of a local bus route failing to tie in with a railway station. Yet, where a formal system possessing arterially is proposed, if anything this appears to be at the expense of accessibility: the so-called 'clear hierarchy' of the Urban Task Force appears to favor a layered hierarchy that amounts to an accessibility constraint, where local buses only go as far as the nearest district centre.
In fact, we can see that one of the problems of the conventional hierarchy is where it may tacitly equate the pedestrian with occupying the lowest rung in the vehicular hierarchy. In doing so, it promotes arterially ahead of accessibility, for pedestrian and vehicle networks alike. Instead, if anything, we could see the pedestrian as occupying the 'lowest' rung in the pedestrian-public transport system. Here, there is no accessibility constraint, and arterially only applies insofar as the pedestrian-accessible network should connect up with some level in tTSu5CKlOs2j1Kg1CUmEt0f1BC3Gmdj9QZuOC53X6eAw=he public transport hierarchy.
For a coherent design strategy, we would therefore attempt to ensure that appropriate combinations oGGU18cQ7EgQ8tfOoq4cE4XbZe7P+fCzMc9PJFenL4kM=f arterially and access constraint were applied in the cases of the different modes. This implies a 'differential' constitution for public transport and the pedestrian, to be reconciled with a constitutional system appropriate for general traffic.
So far we have assumed the conventional case that the tributary constitution, combining arterially with access constraint, is optimal for general traffic. Certainly, at the urban scale, it cannot be denied that for vehicular traffic, it will often be efficient to sacrifice local accessibility (removing intersections, congestion, accidents) for higher standard routes. However, it is possible to argue that arterially is unnecessary, or at least, not a priority, for roads in urban areas. Traffic seems to be able to cope well enough with variations in route standard and route designation within the reality of urban areas.
Together, this implies access constraint without arterially, and the arrangement 'diametrically opposite' that for public transport-pedestrian system. The labyrinth sounds inauspicious, but, it could be argued, this is no worse from the motorist's point of view than the traditional 'differential' case.
Hence, it is therefore possible to suggest a systematic framework for modal fit or utility. As in earlier chapters, this uses simple integer values, and is really just a first suggestion of the direction of the utility function.
The argument for the 'radical interpretation' would be that arterially is not positively required in congested urban areas, where traffic will percolate through the street network as necessary, oblivious to any continuities of strategic significance - unlike public transport, which relies on strategic routes connecting up at interchanges to enable seamless travel. Also, this interpretation does not deny the potential utility for arterially at the national scale. As there is no scope here for a rigoro30539a4e26e3f436c2d5c26fbbc0e849455878aba0f350c7418530773de45ddfus examination of this, however, the conservative interpretation is also retained.
10.2.4 Implied outcomes
Whichever interpretation is made, it is clear that for an optimal fit overall, different network treatments would be required for different modes. This is quite conventional, in that the pedestrian network in suburban layouts is often - but not always – more connective than the tree-like nature of the road layout.
However, the optimal superposition of networks would also imply particular kinds of connectivity between different modes. For example, it implies that pedestrian 'hotspots' and public transport nodes should coincide. The problem with conventional of town centre would be, for example, the sitting of bus stations off on distributor roads away from the central, pedestrian-intensive streets. In such a case, the public transport is equated with (high) vehicular function, not pedestrian function. It also means arterially is targeted such that the main streets can fully real their roles as public transport spines: these being the most direct and continuous routes. The continuity and contiguity of these should be the starting point, rather than in conventional practice where the arterially of the primary distributor network has primacy.
In fact, if we allow the arterially of the vehicular route network to be deprioritised, we could return to the traditional 'differentia pattern for urban structure in general - arterially of the main streets served by public transport, fed by the accessibility of numerous side streets - augmented where necessary by a labyrinthine organization of vehicular routes. Such a labyrinthine case would imply a limited (non-contiguous) selection of vehicle-only relief routes, grade-separated where necessary, with no intersections with local streets or the public realm. This would do as much as is necessary to remove the main flows (or, the stacking of congestion) from the public thoroughfares.
Meanwhile, where vehicular routes were coincident with all-purpose streets, the superposition would still imply that vehicular connections to side streets would be limited, even as pedestrian connections were maintained, as in the traditional case.
This could accord with a kind of 'string-of-high-streets' model, where major routes (e.g., radials) with high public transport and pedestrian function are coincident with commercial frontages. This form would be like a linear extension of the 'characteristic' form, focusing on a public transport spine. However, the pedestrian access streets would form a connective mesh of grid streets, giving accessibility in all directions, rather than 'focusing' on public transport stops (as with beads-on-a-string).
6c1B4D41dkRz9rzmdDztuRqwt72LJxXKfbDXEocgv2k=Such a scenario would be commensurate with a tangible prioritization of pedestrian and public transport, where vehicular traffic is accommodated as a second priority.
The combination of 'differential' for pedestrian and public transport, and 'labyrinthine' for general traffic, may be more physically practicable than the combination of differential plus tributary (attempting to optimize for all modes). Indeed, the traditional plus labyrinthine seems like a realistic balance, as indeed it reflects much of the reality (if not the ideal) of existing urban areas, with their mix of high streets, resurgent with pedestrian and public transport priority, and punctuated by the occasional relief road.
We need not specify in detail what kinds of configuration or composition such a constitutional organization would entail. One could point to de facto examples of such a traditional labyrinthine arrangement in today's urban areas. But the whole point is that the layouts would flexibly arise from local circumstances; the routes and patterns would, as Buchanan would say, 'work themselves out'.
What can and shall be done here, however, is to link these constitutional principles with the configurational considerations of preceding chapters, drawing together the set of design components from the thesis as a whole into an integrated framework.
10.3 Towards an integrated design approach
What is needed are theories.., that are as nonspecific as possible to particular solutions in the generative phases of design in order to leave the solution field as large and dense as possible, and as specific and rigorous as possible in the predictive phases in order to be able to deal predictive with unknown forms where the need for effective prediction is greatest.
The drawbacks of many normative theories of design is that they are over-specific where they should be permissive, and vague where they should be precise. Bearing this in mind, the proposed approach to design should be kept as flexible as possible, regarding its synthetic components.
The major work of the thesis has been in making explicit the relationships between different transport modes, network structures and morphogenetic processes. It is felt that the freshly explicit nature of these relationships empowers the designer to make intelligent choices about design, appropriate to individual circumstances.
Accordingly, it is not considered helpful to end here by narrowing down the potential solution space to a fixed set of specific design recommendations. Hence, no single design strategy is pursued; no particular 'preferred' or 'discouraged' layouts proffered. While some suggestions for the content of the individual design components are given, these are illustrative, and not intended as hard and fast rules. Rather, the main task of this section is to synthesize the main components - expressed in sufficient detail to explain their c19e433418a57f03a3eed400b54dac6f2ontent - to show how they might be put to work together.
The three main aspects of design considered here are (1) modal function; (2) configuration; and (3) constitution. This section first reviews or elaborates the main components of each of these before drawing these together into an Integrated Design Framework. This is followed by reflections on the implications for the design debate .
10.3.1 Transport mode considerations
Modal fbe771972946a811e7ebb706b50d5795cunction can represent a starting point for design. This thesis has concentrated on three; in practice, all relevant modes would be considered.
In one sense this represents the separate consideration of different modes, in that their individual characteristics are explicitly accounted for. That said, the separate considerations of the different modes are then combined in a single integrated approach.
This differs from approaches of the past in which private motor traffic was prioritized as far as the road system was concerned, and other modes were only treated in specialized contexts, often to different rules.
For example, public transport might either be considered completely separately from road network form, or be lumped in with general vehicular traffic. In the Buchanan Report, buses are supposed to use the distributor roads; however the distributor road system seems heavily geared to the needs of the individual motor traffic.
The Buchanan Report recognized pedestrian use of access roads, and also envisaged a separate footpath system, but this separate footpath system was not explicitly linked into the road hierarchy.
10.3.2 Configurational components
The configurational components have been introduced and discussed individually from Chapters 7 to 9. These components are linked to each other and also to the modal and constitutional aspects. For example, the positions of routes on the routegram may be equated to some degree with modal function, and hence potential role in a hierarchy of functional types. Within the integrated framework, a variety of possible approaches to design and analysis are possible. One possible chain of relationships is illustrated. This represents an analytic path within the design 9bbfebb0db3d3301814dc634a50a5887process, reflecting the order in which the components have been introduced in the thesis.
10.3.3 Constitutional components
The possibilities for a constitutional approach have been outlined in principle. This section suggests in more detail what these components would entail. These comprise52afdfe50240063f5d1e6f9f722d95a1 (1) hierarchy of route types; (2) allowable connections between types; (3) allowable types of connection; and (4) a means of combining these in a single construct: the 'constitutional archetype'.
1. Hierarchy of route types
The need for a hierarchy is to help provide some degree of control of relationships between route types, even as some conventionally imposed constraints are released. The suggested matrix gave separate consideration to pedestrians and public transport as well as general vehicular traffic, and to streets (i.e., routes with frontages), then combined these in a single matrix. Here, a set of 18 recognizable route types is considered – distinguished here by modal function but not frontage function.
More route types or permutations could be generated as desired. For each type a set of standards could be specified, e.g., construction standards, widths, parking and loading control, frontage access, permitted frontage development and uses, etc. These route types can be seen to tie transport, built form and land use considerations together. They represent the basic building blocks of urban structure.
This 'hierarchy' effectively presents the allowable coincidence of modes along particular types of route, but not yet the allowable connections between these.
2. Allowable connections between types
A suggested system for rules of connection between different route types. Although the table may look complex as formally expressed, it can be handled quite intuitively once its logic is grasped: it is simply based on the two principles of arterially and access constraint discussed.
3. Allowable types of junction connection
The allowable connections between route type have just been suggested. In practice, the types of connection would also be regulated. The determination of appropriate junction type is often a detailed technical exercise in its own right. The point here is to consider junction type purely in terms of nodality, as a contribution to overall network connectivity.
In a purely configurational approach, connectivity could be targeted by specifying some kind of type or network property. However, it is also possible to influence network connectivity at the constitutional level, by specifying allowable junction nodality. Networks with 4-way or 5-way junctions appearing in more circumstances (e.g., more combinations of route type intersection) will tend to be more connective than those where multi-way junctions are discouraged.
Here, types of vehicular junction are considered explicitly (It is assumed that pedestrian connections are not statutorily limited, and that public transport connection constraints would follow those of vehicular routes.)
Illustration of the possible content of such a table. The matrix illustrated would imply and require that any 4-way junction on a v3 or v=4 route would have an appropriate form of junction control, such as traffic signals. Traffic signals represent a traditioncd3e5b3c3fc4b6db23e7793106dd180cal and urban solution - recall Alexander's traffic light and newsrack as opposed to the free-flowing highways and underpasses of conventional layouts. Urbanity and safety could both be accommodated - as they are in all existing cases where crossroads and 5-way junctions continue to function in traditional urban areas, where the single-minded pursuit of traffic fluidity had not caused them to be swept away.
Where, when and how 4-way and 5-way vehicular junctions were introduced would lie outside the scope of the thesis.
The point here is that the specification of different junction types (defined by nodality) can be made here as an explicit component of the constitutional program for urban structure. Through such local rules, network connectivity can be influenced, without necessarily prescribing the connectivity at the network level.
By making a relatively small number of changes to the program (e.g., allowing 4- way junctions to be formed by two routes, with appropriate signalisation) we would be able to create a much expanded 'solution space' of allowable patterns.
4. Constitutional archetype
The last three items (hierarchy of types, allowable connections and junction type) may be collated together in a single graphical construct. This might be called a 'constitutional archetype' since it represents a basic theoretical structure which may be transformed into any number of individual configurations by expressing (selecting) or suppressing different components in different proportions and sequences.
The constitutional archetype could be used as a basis for design guidance, route types according to modal function, but ultimately frontage function and other land uses could be included, filling in what are currently blank interstitial blocks.
The constitutional archetype shown here is effectively a more complex evolution, which may also be interpreted as a simple constitutional archetype, albeit based on a more limited number of route types and allowable connections. The present proposal therefore has the conventional case embedded within it. In both cases the configuration adopted to illustrate the allowable constitutional relationships does not imply a recommended literal configuration.
This constitutional archetype may be contrasted with a configurational archetype which would be one which represented the recommended configuration, which would then be transformed to real site compositions by the use of dimensioning sets for distances and angularity. Steadman's archetypal building would be a 'configurational' archetype.
The principal components from their key relationships. In some cases the relationships are envisaged to pass in a specific direction; in other cases the relationship may go either way: in some cases, this will equate with the difference between design synthesis and analysis.
Within this framework, a variety of approaches to design and analysis is therefore possible. These will all involve to a greater or lesser extent some kind of iterative process, as the addition of any routes will affect the character of intersecting routes and the network as a whole, as these are adjusted with target properties in mind.
For any of the approaches in (1) could represent the basis for national or institutional design guidance, while column (2) could represent an assumed or encouraged convention.
For example, a configurational approach could be introduced, targeted towards a neo-traditional outcome, by recommending target network types or connectivity values. Here, the inter linkages between network type, route type and modal affinity, as a more sophisticated system than a simple 'preferred'/'discouraged' polarization.
Alternatively, on the constitutional side, existing design guidance could be modified by the incorporation of any or all of the relevant design components. For example: (a) the expansion of route typology (e.g., from the present handful of types to 25 types); (b) the application of different constitutional organizations to different modes; or (c) the expanded use of 4-way junctions could each be incorporated independently of each other. Their detailed content could be more - or less - conservative than the suggestions given in this section. And since the Integrated Design Framework accommodates the conventional approaches, it could be adopted in principle prior to making any changes to the individual design rules or recommendations.
10.3.5 Implications for the design debate
It is finally possible to compare the characteristics of four particular approaches which are of relevance to the design debate. These expressed as combinations of different design approaches arising in different layout outcomes.
Case (c) is the theoretical conventional case, which is in principle constitutional, in which a tributary hierarchy is promoted and the use of crossroads discouraged. However, in practice we can also recognize (a), the apparent de facto conventional case, in which the tributary constitution is commonly associated with tree-like configurations (Buchanan's Tree), giving the impression of the configurational replication of standardized loop-and cul- de-sac layouts.
Case (a) represents the 'problem': this is the kind of layout for which engineers are blamed, and which neo-traditional urbanists would like to replace. Their solution often appears like case (b), in which conventional hierarchy is rejected, to be replaced by a pattern-based' approach, with grid-like configurational or compositional exemplars.
However, the argument of this thesis has led to the suggestion that a constitutional approach may be a more appropriate solution: this allows a variety of route types and superposition of hierarchical organisations, and more liberal forms of connectivity, yet the final form is not prescribed. This is effectively a kind of neo-organic approach (in spirit, more akin to the organic approach of Chapter 2 than the post modernistic neo-traditional case). It seems to combine the advantages of the flexibility of constitution-led design with the possibility for grid-like (as well as tree-like) layout outcomes. Indeed, the latter advantage can be seen as a direct consequence of the former.
This can be demonstrated in terms of the potential breadth and depth of the solution space. Neo-traditional design by configurational or compositional exemplar could be described as being 'broad but shallow' in that a relatively broad range of patterns is allowable (eg, grids as well as loops and cul-de-sac), but the superficial level at which the patterns are reproduced theoretically limits the potential solution space. Using a constitutional approach can theoretically deepen the solution space, broadening it in the process.
However, in practice, in the de facto conventional case, this range of theoretical possibilities is pruned: firstly by explicit constraints such as the rejection of the crossroads and tributary hierarchy (which prevents minor roads from directly accessing major roads), and secondly, by the implicit association of tree-like configurations with a tributary constitution (Buchanan's Tree). The de facto conventional case is therefore limited to a 'deep but narrow' slice of the solution space. It therefore appears to be more standardized than the neo-traditional menu - since many 'trees' look the same (although the potential combinatorial volume of technically distinct, if practically indistinguishable, tree-like layouts is theoretically vast).
A more permissive rule system, allowing more connective layouts, could realise QixAMZPgMPRvpdwB3jWxsA==the full potential gain in depth and breadth of a constitutional approach, whose solution space could potentially include both the favored neo-traditional layouts and conventional layouts, as well as other, non-preconceived forms.
The constitution-led approach amounts to a bottom-up method of design that is the antithesis of Alexander's Tree. Where we combine the constitutional synthetic approach with a component of configurational analysis, in an iterative procedure of design then we come closest to satisfying Hillier's criteria of permissive creation and precise diagnosis quoted at the start of this section. And, as far as the design debate is concerned, we can also reconcile the favored qualities of two opposing approaches - the robust rationality of the engineering method and the variety and organicism envisioned by the urbanist.
10.4 Chapter Discussion
The approach to the design of urban structure developed in this Chapter is based on local rules of structural connection of the transport network. This approach is constitution-led, rather than configuration-led; or program-based rather than pattern-based. This allows urban structure to grow in an 'organic' way that can lead to unanticipated forms. This avoids both the 'monolithic modernism' of the master plan, and the more prescriptive pattern-based tendencies of neo-traditional urbanism. Yet it can lead to desired forms which emulate the semi-lattice patterns of traditional urbanism. In effect, it allows the generation of structures possessing 'coherence', striking a balance between total chaos and total conformity.
This approach therefore reflects ways in which traditional characteristic structure has been formed. It is commensurate with an evolutionary mode of urban creation, driven by entropic probability. (D'Arcy Thompson notes that 'entropy' is a literal translation of 'evolution' into Greek; 194811). In doing so, it goes some way to fulfilling Michael ‘s call for a new approach, quoted at the outset of the chapter.
We can see how characteristic urban structure can be (re)created, where coherence is emergent, without need for a 'blueprint'. To the extent that city design may be compared with organic creation, this has a resonance with the bottom-up processes of 'artificial life', where viable features emerge in situ. This could be contrasted metaphorically with a kind of top-down 'Frankenstein' approach to creating life, a crude assembly of imported parts.
Of course, the bottom-up approach cannot rest on philosophical appeal, but must be functional in urban terms. A potential problem of top-down master-planning is that it runs the risk of being dysfunctional - a 'hopeful monster' gone wrong. This would especially apply in the case of large-scale intervention into existing urban fabric, and the speculative leap to novel, grandiose forms, such as Le Corbusier's Plan Voisin for Paris, or indeed Buchanan's Fitzrovia.
In contrast, an urban structural approach is suited to a more gradual, evolutionary process, which is consonant with generating organized complexity, based on incremental contiguity. This incremental contiguity is in the nature of urban structure. The contiguity is a feature of transport network structure, that is not possessed by discrete buildings or districts. The incrementality is a feature of the growth of urban settlements, that is not normally a feature of other designed objects.
According to Ebenezer Howard:
Insofar as 'completeness' connotes viability at all stages of growth, the application to settlements seems reasonable. However, the passage also seems to connote the idea that a town as a whole ought to be a consciously fashioned product: the argument for town planning.
This thesis has not found the need for such kind of 'creationist' planning when it comes to urban structure. A building or engineering structure is quite properly 'creationist', where the whole depends on the parts being put together in a certain supportive manner. But a city need not be creationist: it can be quite naturally evolutionary.
In fact, Howard himself contrasted the construction of a town and the building of a bridge, stating that 'a bridge is not a bridge until the last rivet is driven home'.
His point concerned advance capital; the point here is that planning the construction and overall form of a bridge is essential if it is to be functional, but this is not necessarily the case with the structure of a settlement.
In the end, perhaps the evolutionary, bottom-up approach to urban structure suggests not a planner-God, nor a 'highway architect', but some kind of urban-structural engineer. Perhaps the attempt to design whole cities is barking up the wrong tree. Perhaps it is sufficient to concentrate on the constitution of the parts. Perhaps, as Mies van de Rohe put it, God is in the details...