New Urbanism movement has arisen since the 1980s. It promotes environmentally friendly habits by creating walkable neighborhoods containing a wide range of housing and job types (Boeing et al., 2014). The principles of New Urbanism, such as walkability and pedestrian-friendly community, diverse neighborhoods and smart trans- portation, encourage the re-establishment of mixed-use and sustainable urban development. As a representative approach of New Urbanism, Form-Based Code is increasingly concerned due to its creativity in zoning academia and practice.

Form-Based Code is a land development regulation that fosters predictable built results and a high-quality public space by using the physical form as the organizing principle for the code (Institute, 2014). In the literature dealing with definition of Form-Based Code, plenty of studies focused on comparing this novel approach with conventional zoning (Parolek et al., 2008). Conventional zoning describes zoning and urban regualtion by the authority which designates legal areas in a municipality to permit and prohibit land use (Lemar, 2015). In the procedure of zoning development, land use is the primary consideration and foundation. It stems from the emergence of conventional zoning during the early twentieth century. At that time, the goal of public health, safety, and welfare drove zoning to segregate residential districts from industry districts (Kan, 2012). However, with the built municipality increased rapidly, the limitation of conventional zoning emerged. There were problems of over zoning which left the cities those spotty areas which aren’t quite residential or commercial (Frost, 1958). In the 1980s, Form-Based Code started to work in urban projects as a critical response to conventional zoning. Just as its name implies, Form-Based Code regards form as the essential consideration rather than land use. Form-Based Code is different from conventional zoning in terms of the process by which they are prepared, the substance of the standards they contain, the mechanisms by which they are implemented, and the built form they produce (Parolek et al., 2008). It is argued that contemporary urban developing flow is a continuous mixture. Form-Based Code that promoted mix-used and continuous growth seems more appropriate to cities today.

One representative Form-Based Code project is the Seaside Code of Florida which developed by Duany Plater-Zyberk & Company. The Seaside Master Plan of Form-Based Code regulated the buildings to ensure the creation of an urban environment similar to that of a small Southern town of the period before 1940 (Plater-Zyberk & Company, 1980). This project reflects a core value of Form-Based Code – revival rather than newly-created. Inniss called Form-Based Code as back to the future (Inniss, 2007) due to its idea of protecting urban morphology and space texture. According to Borys’ Code Study, there are more than six hundred Form-Based Code projects adopted or in the processing as of February 2017. As an American zoning approach, most of the Form-Based Codes are clustered in the USA. If Form-Based Code has the capability to positively influent on the development of cities and peripheries, we hypothesis that it can bring similar benefits to other areas out of the USA. The research attempts to evaluate Form-Based Code in Tsim Sha Tsui area of Hong Kong in response to the hypothesis.

It is necessary to understand Form-Based Code’s zoning method before experimenting with the evaluations. Conventional zoning subdivides land into a series of zones such as commercial or residential and then describe the regulations of each zone through tables and images. Differently, Form-Based Code subdivides “transect” into various zones as Tn (n=1, 2, 3 . . . ). It generates regulations and codes through two-dimensional and three-dimensional visualized languages for each zone. In Form-Based Code, transect matrix provides zone types with specific forms. The concept of transect matrix is originally used by naturalists to describe the characteristics of ecosystems and the transition from one ecosystem to another. Duany has applied this concept to human settlements, and since about 2000 this idea has permeated the thinking of new urbanists [(Zhang & Schnabel, 2017); (Congress for the New Urbanism, 2016)]. The transect is described as:

“. . . it arranges in useful order the elements of urbanism by classifying them from rural to urban. Every urban element finds a place within its continuum. For example, a street is more urban than a road, a curb more urban than a swale, a brick wall more urban than a wooden one, and an alley of trees more urban than a cluster. Even the character of streetlights can be assigned in the transect according to the fabrication from cast iron (most urban), extruded pipe, or wood post (most rural)” (Steuteville, 2009).

Figure 1 presents a standard transect matrix of Form-Based Code from natural zone to the special district, which defines the hierarchical development scales (Bum Kim et al., 2011). T1 represents natural zone without any artificial constructions. T2 to T6 describes a continuous improvement of the artificial degree from rural to urban core zone. Based on the determining of zone types, models can be generated to describe Form-Based Code at the conceptual stage of zoning.

Form-Based Code utilizes illustrations and models to describe the zoning intentions, logic and codes. Directly visualized models contribute to the prediction of zoning results. This section provides an attempt on experimenting Form-Based Code in a high-density environment by using Tsim Sha Tsui, Hong Kong as the experimental site. The attempt is supported by an extended transect matrix, a parametric modelling system and a series of Form-Based Code models.

Using the parametric model that developed in Grasshopper 3D, this section evaluates the connectivity of the Form-Based Code modelling results. Connectivity optimizing works for examining the public space usage and predicting movement routes under Form-Based Code intention in the volumetric morphology. Scenarios evaluations are conducted in the perspectives of street-scale pedestrian movement, underground open space usage and volumetric walkable routes.

In the aspect of street-scale movement, the possible pedestrian flow is simulated by controlling the building coverage rates, setbacks and walkable areas. The elements influencing street-scale regulation are manifold, but walkability is increasingly valued at street canyon and neighbourhood (Zhang & Schnabel, 2017) especially in high-density cities. As Southworth has presented, not only does pedestrian transportation reduce congestion and have a low environmental impact, it has social and recreational value (Zhang & Schnabel, 2017). He argues that six criteria for the design of a successful pedestrian network: connectivity; linage with other modes; finely grained land use pattern; safety; quality of path; and path context (Southworth, 2005). In Form-Based Code models, the connectivity is evaluated through parametric thinkings. The research transfers the perceive pedestrian space into the parametric model and analyse the pedestrian movement in algorithm platform.

The component Galapagos of Grasshopper 3D works as the instruments to evaluate connectivity. It uses an evolutionary-based solver to calculate solutions. Using one street-scale site of Tsim Sha Tsui as an example. The site contains part of zone T6-1 and T6-4. A script is written in Grasshopper 3D with the main component Galapagos. Galapagos has two connecting points, Genome and Fitness. They are both for importing information. The Genome is to connect with independent variables. Coordinates of each block or plot are the independent variables, and the length of the connections between the nodes is the part that should be measured through Galapagos (Figure 8).

The parametric script for optimizing connectivity is described in Figure 9. If assuming the buildings in contexts are regular symmetrical rectangles, they can be simplified as parametric nodes in the modelling process. Through running the Galapagos component, the pedestrians’ activity path can be roughly simulated and measured. In the high-density areas of Tsim Sha Tsui, shorter walkable connections between buildings contribute to a convenient and effective pedestrian system. While longer walking paths offer more opportunities for stores being noticed by customers especially in a place with mixed functions of commercial and residential. Galapagos in Grasshopper 3D helps to calculate an appropriate walking distance for pedestrians. When processing the components, Galapagos offers a pop-up screen for changing the options (Figure 10). Form-Based Code team can choose the Fitness type, maximize or minimize, to control the processing results.

According to the setting of the Fitness type, the pop-up screen of Galapagos Solvers provides four data-analysing parts (Figure 11). The yellow line chart on the top of the screen reflects the tolerance degree. When the data-line of right side becomes stable, the value is nearly the optimum solution. For example, if the Fitness is set as Maximize, the model presents a relatively maximize routes when the data-line is stable at one value. In the underneath part, the point and line figures represent the measures of variability; the numbers represent the tolerance values (Zhang & Schnabel, 2017).

For underground level, the space usage is simulated according to the transect zone type subdivision and existing underground exploitation. Based on the local Government’s guidance, the underground pedestrian corridors of Tsim Sha Tsui will be extended from the current Tsim Sha Tsui Metro – Tsim Sha Tsui Metro East system to the Tsim Sha Tsui Metro – Tsim Sha Tsui East – Guangdong Road system. The underground area under Kowloon Park has the priority on the project lists. According to the zoning concept of Form-Based Code, the underground Kowloon Park belongs to the zone in approximate rectangle shape. The paper uses the undergroudn zone of Kowloon Park as a modelling example to test the possible space usage.

Assuming there are random nodes in the underground site, each node represents a specific function component, such as underground shops, gathering points, public art show, or toilets. Since no subway transfer requirement in this zone, the use of underground space is mainly for extending open space, creating a social place and duplicating walkable streets. Programming the nodes by parametric script, the relatively appropriate locations of the functions can be decided taking advantage of Solvers (Figure 12). Through programming in Grasshopper, the connectivity between nodes can be measured by Maximum or Minimum Fitness. The results help Form-Based Code to decide the plan of space usage.

Vertical movement convenience works to measure the volumetric space usage. The parametric platform allows Form-Based Code team to simulate people’s activities in the vertical direction and calculate relatively short conversion paths. The programming in Grasshopper 3D provides possible vertical routes both inside and outside buildings (Figure 13). The red lines represent walking paths. Red boards are simulated connections inside or between buildings. Through Galapagos, it can be calculated which routes are shortest between different function nodes. For releasing large pedestrian flow in the vertical direction, shorter path means more convenient transportation. Main approaches to practice the vertical transportation are elevators, stairs and slopes.

Manipulating Galapagos works to simulate node connectivities in the parametric modelling at the conceptual stage of Form-Based Code. Possible zoning results are simulated by flexible manipulation of parameters and components. Further details of constructing vertical connections and underground space need deeper analyse in downstream architecture and urban design.

This paper provides a demonstration that parametric methodology has the technical capability to support Form-Based Code evaluation in a high-density environment. Using Tsim Sha Tsui, Hong Kong as the experimental site, the parametric Form-Based Code provides an alternative approach to zoning.

Form-Based Code usually uses graphic language to describe urban regulations and provisions. It converts texts and tables to images. Although the generation methodology is different from conventional zoning, the original Form-Based Code mainly relies on manually drawn. Generation and modification are cumbersome especially in large-scale urban projects. For real implementations, it is common that different variables are related to each other to define a place. Modifying one variable causes changes to other variables. Manually modification increases workload. Introducing parametric thinking makes the generation and modification process more automated. Form- Based Code team can receive a series of performance feedbacks by changing the target parameters. In addition, parametric modelling instruments, such as Grasshopper 3D, provides evaluation-related components to automatically calculate extremum value. Programming these components, the software simulates activity routes between geometries. Parametric thinking facilitates Form-Based Code by integrating generation, modification and evaluation in one modelling process.

Form-Based Code makes it possible to generate mix-used urban space through weakening land-use division and strengthening form partition. When attempting Form-Based Code in Tsim Sha Tsui, it appears to be an effective and viable option to predict zoning results before implementing in real projects. Form-Based Code is related to floor area ratio, average pavement width, road hierarchy, infrastructure, the depth-width ratio of streets, and max height. Defining the requirements of these variables, a series of zone types are generated to describe the unique space characteristics of the experimental site. With an expected physical outcome, Form-Based Code provides models about what the future urban space looks like. Connectivity is a basic requirement for volumetric cities like Hong Kong in terms of high-density transportation flow, block permeability, and diverse functional areas. Hong Kong already has a highly-developed connective urban system. This paper employs parametric simulation to evaluate the connectivity of Form-Based Code models. The evaluation results provide possible spatial streamlines in the volumetric morphology. The connectivity evaluation supports the enhancement of the pedestrian connection to improve the walking environment performance.

This research creatively conducts an experiment that examining Form-Based Code in a volumetric morphology of high-density cities in Asia. There is clearly more explorations and deliberations to be done in the area of Form-Based Code analysis, such as the cooperation between Form-Based Code and established zoning system. We can conclude that Form-Based Code has potential to implement effectively in high-density cities, supporting local urban planning and design departments, developers, and householders’ efforts to address the challenges of high-density development and build an amenity living place.