- Email: [email protected]
Reconstructing urban contexts online for interactive urban designs
Reconstructing urban contexts online for interactive urban designs Chengzhi Peng and Peter Blundell Jones, School of Architecture, The University of Sheffield, Arts Tower, Western Bank, Sheffield S10 2TN, UK The Sheffield Urban Contextual Databank (SUCoD) system was developed as a Web-based urban information system to host the urban data gathered by the Sheffield Urban Study programme. Our development of the SUCoD system provides user-centred dynamic retrieval and visualisation facilities for accessing multidimensional city datasets. The paper reports on a study of applying the SUCoD resources in an educational setting to explore its usability in two accounts: presenting urban site investigations as online urban narratives and the development of 3D schematic urban proposals. The current experiment suggests that SUCoD could be further developed into an online Urban Design Collaboratory. 쎻 c 2003 Elsevier Ltd. All rights reserved. Keywords: design education, urban design, modelling, virtual reality, system design
T
he necessity of investigating and understanding an urban context in which an architectural design is to be developed has been constantly emphasised by civic authorities, design educators, critics and practitioners. At most schools of architecture, we can see that studio design projects often start with some form of contextual investigations of the city sites. The contextual study involves, typically, the gathering of information on the physical, historical and social aspects of an urban site, and the production of various kinds of records such as site drawings and models. Students then develop their proposals in relation to these contextual records. The results may turn out drastically in contrast with the existing conditions of the site or they may appear much blended into the urban setting. From a pedagogical viewpoint, one could argue that architectural design can never be entirely ‘site-determined’, yet a general consensus seems to www.elsevier.com/locate/destud 0142-694X $ - see front matter Design Studies 25 (2004) 175–192 doi:10.1016/j.destud.2003.10.005 2003 Elsevier Ltd All rights reserved Printed in Great Britain
175
1 Peng, C, Chang, D, Blundell Jones, P and Lawson, B ‘Exploring urban history and space online: design of the virtual Sheffield application’ Design Studies Vol 23 No 5 (2002) 437–453 2 Peng, C, Chang, D, Blundell Jones, P and Lawson, B ‘Making Virtual Cities Dynamic’ International Journal of Design Computing, Special Issue on Designing Virtual Worlds Vol 4, (2002) (http://www.arch.usyd.edu.au/kcdc/ journal/vol4/index.html) 3 Peng, C, Chang, D, Blundell Jones, P and Lawson, B ‘Dynamic retrievals in an urban contextual databank system using Java-CGI communications’ in B de Vries, J van Leeuwen and H Achten (eds) Proceedings of Computer Aided Architectural Design Futures 2001, 8–11 July, Eindhoven University of Technology, Eindhoven, The Netherlands, Kluwer Academic Publishers, Dordrecht (2001) pp 89–102 4 Ennis, G, Lindsay, M and Grant, M ‘VRML possibilities: the evolution of the Glasgow model’ in Proceedings of the Fifth International Conference on Virtual Systems and Multimedia (VSMM ‘99), Dundee, Scotland (1999) 5 McCullough, M and Hoinkes, R ‘Dynamic data sets as collaboration in urban design’ in M Tan and R Teh (eds) Proceedings of Computer Aided Architectural Design Futures 1995, Centre for Advanced Studies in Architecture, Department of Architecture, National University of Singapore, 24–26 September, Singapore (1995) 6 Batty, M, Dodge, M and Jiang, B ‘GIS and urban design’ in S Geertman, S Openshaw and J Stillwell (eds) Geographical Information and Planning: European Perspectives, Springer-Verlag, Berlin (1999) pp 43–65 7 Pietsch, S, Radford, A and Woodbury, R ‘Making and using a city model: Adelaide, Australia’ in Proceedings of the 19th eCAADe Conference, 29–31 August, Helsinki (2001) pp 442–447
176
subscribe to the objective of fostering the students’ appetites or inspirations for engaging in ‘dialogues’ with the urban contexts in question. Based on a recent experiment in an undergraduate urban design computing course, this paper reports a study on the usability and usefulness of a Web-based dynamic virtual city information system for directly supporting urban investigation and 3D urban design. Since late 1999, we have been developing the Sheffield Urban Contextual Databank (SUCoD) system as a Web-based open-source repository to host historical records of the city of Sheffield.1–3 We are certainly not the first to explore how digital technologies could be applied to urban studies and design. In early 1980s, the ABACUS group at the University of Strathclyde pioneered 3D computer modelling technologies for visual simulation of the city of Glasgow at a large scale, which have generated numerous educational and industrial applications.4 Working in the combined architecture and landscape domain, McCullough and Hoinkes5 considered that central to urban design are ‘dynamic, multivalent datasets for collaborative processes’, which could be better supported through a rich urban information gathering and interaction environment. Similarly, collaborating with graduate urban design students at the Oxford Brookes University, Michael Batty and co-workers6] investigated how geographical information systems (GIS) could be linked with computer-aided design (CAD) to provide intuitive GIS-driven visualisation and communication facilities for participative urban design. More recently, Susan Pietsch and colleagues7 reported their studies of using sections of the city models built for the city of Adelaide in student design projects. As seen from the related studies, the creation of urban contextual resources at a large scale and applying the digital resources in architectural and urban design still present considerable theoretical and technical challenges.
1
Studying the historical city fabric of Sheffield
The ‘Sheffield Urban Study’ programme was first launched in 1998 at the School of Architecture, University of Sheffield. The starting point for the Urban Study was a consensus among the staff that context is important, and that one should always know the history of a site in order to understand how it was formed and what it means, whatever use is eventually made of that information. By accessing and studying the historical maps and records from the City Archives, Local Studies Centre and the Hawley Collection, students at the Diploma School during 1998–2001 built a physical model of the city as it stood in the year 1900, at a scale of 1:500. This was accomplished by dividing the city with a grid into 200 m2, each of which was investigated and modelled by a group of four or five students. They were asked to report on the ages and uses of all the buildings at the
Design Studies Vol 25 No. 2 March 2004
Figure 1 A north-south aerial view of the assembled Sheffield 1900 model
period under study. Each student was also asked to choose and study a particular building, comparing it with the equivalent building today if appropriate.8 To date, a total of 41 unit models have been constructed, revealing a view of historic Sheffield unseen before (Figure 1). Since its inception, the urban study archive has grown progressively following the urban grid as an analytical framework which can be extended at anytime to cover other areas of the city. To the research community, we believe that the Sheffield Urban Study resources have the following potential values:
8 Blundell Jones, P, Williams, A and Lintonbon, J ‘The Sheffield urban study project’ Architectural Research Quarterly Vol 3 No 3 (1999) 235–244
앫 Heritage. The city of Sheffield grew from little more than a village to one of the largest cities in Britain between 1750 and 1900 because of its industries and the world-famous Sheffield steel. The heavy steel industry continues in mechanised form far outside the city, but the tool and cutlery industries have almost faded away through international competition. Though initiated by it, the city has been slow to appreciate the value of its industrial heritage. The industrial installations and buildings, usually in very poor condition at the point of economic collapse, have very largely been razed to the ground. Even the street pattern has in many cases been abandoned. Yet from a heritage point of view, Shef-
Reconstructing urban contexts
177
앫
앫
앫
앫
field could be of great international interest, and could do much more to celebrate its origins and industrial heroes like Benjamin Huntsman, inventor of the crucible process. Every year, however, more industrial remains are destroyed. The Sheffield Urban Study programme systematically identifies the history of every site, giving a necessary overview of the layers of the city and their relative importance. Current planning and understanding of context. Developers and architects tend to regard a site as a blank slate unless buildings on it are listed, yet often it is possible to preserve some characteristics or even parts of existing buildings within a new design to give it some historical continuity. With a database such as can be provided through the Urban Study, the evidence is at least quickly to hand and the possibility can be considered. Student experience. As an educational course for architectural students, the Urban Study engenders a personal awareness of quality of place and of the layers left by history, so never again can a site be regarded as ‘a nowhere place’. Involvement in compiling the database teaches students techniques of recording and analysis, the use of sources such as archives, and the kind of thing to be sought and found in a typical urban site. It is a necessary training for those whose responsibilities include some duty to preserve as well as to change the nature of places. Database. The Urban Study produces a unique record of the city, which helps in understanding its form and social history. With an increase in resources, this could be made available to a much wider circle of users. A generic methodology. The Urban Study has pioneered new techniques in historical analysis and recording, both in its initial paper-based and physical modelling phase, and through its translation into an electronic version. In essence, the methodology can be re-applied to any other city, and not just industrial ones. Although it might initially appear straightforward, the translation into electronic form has revealed many difficulties and required considerable reflection about methods of storing and classifying information, and about the kinds of information that lend themselves to electronic media. In some areas, they have great advantages over paper and in others they are lacking. The greatest challenge has been to create classification methods that remain to some extent open in order to embrace types of data as yet unforeseen.
2
The Sheffield Urban Contextual Databank (SUCoD) system
As it is, the Sheffield 1900 database has taken the form of a paper file and a cardboard model of each grid square. The ‘box-file’ of documents for every grid square has proved a most practical way of assembling data
178
Design Studies Vol 25 No. 2 March 2004
initially. However, we find that uses of the physical state of the data are limited in several aspects: (a) The contextual data acquired in its current formats do not allow for efficient storage, transportation, management, and rapid reproduction. Clearly, the physical state of the data does not facilitate more widely collaborative research and debate on proposing new architectural interventions in an educational or professional setting. Also, the graphic conventions and formats used are not altogether consistent. (b)Urban contextual study is potentially a continuous process undertaken by academic or professional design research groups, and material for the database may be accumulated over years. A physical approach to modelling and documentation has limited flexibility for extending a database consistently over a long period. (c) A physical approach can follow a clear organisation if parts of the contextual study are geographically tightly connected as with the Sheffield 1900 project, but this is not always the case. Contextual inquiries may involve a region with the parts geographically dispersed, and with no imminent connections to be made. In such cases, it is difficult to establish the organisation through direct physical modelling. (d)The aim of understanding and showing how places have evolved over a period of time can be achieved through juxtaposition of a series of maps, drawings or models. To display the evolutions three-dimensionally in physical models, however, is cumbersome if not impossible. Simply replacing parts of the model does not offer optical continuity or visual cues for appreciating processes as they change.
Parallel to the Sheffield Urban Study project, we have been investigating how to organise the urban contextual information electronically, taking into account the purposes and requirements of the study as already discussed. Our aim is to establish an effective methodology for creating and managing multidimensional urban datasets that could support our design research and education programmes. We also take into account the unique requirements of Sheffield as a post-industrial city and the potential usefulness of research outcomes. Over the past two years, we have designed and implemented the Sheffield Urban Contextual Databank (SUCoD) system as a pilot study of digitising the urban resources. SUCoD is able to deliver multidimensional urban datasets such as interactive maps and virtual reality models by responding to users’ choices and requests transmitted over the Web
Reconstructing urban contexts
179
(http://sucod.shef.ac.uk). In developing the system, we were particularly interested in experimenting with an alternative methodology that will lift some of the limitations seen in the existing approaches to building ‘virtual cities’, which have been largely based on static urban modelling. In our view, existing methods of making 3D digital city models pose fundamental difficulties in making these models applicable to directly supporting contextual urban study, planning, and design. We commonly see that the basic recognition and definition of an urban context often varies from one individual to another. An architect, for instance, may take the entire city or town as the contextual framework to work with when designing a primary school for a certain location, while another architect may choose to focus on a specific cluster of buildings surrounding the site as the sole context to be addressed. We believe that it is essential for an urban information system to allow the exercise of individual differences in recognising and defining the focal locations and boundaries of an urban context, as there lies the origin of creativity that may contribute to qualities of urban design in a fundamental way. In its current release, SUCoD provides users with map-based interface to select and retrieve the following kinds of urban contextual resources (Figure 2). Scanned historical maps. Historical maps are of an important type of urban contextual resource that users may read with a great interest. To enable
Figure 2 The (SUCoD) system on the Web as seen through Netscape
180
Design Studies Vol 25 No. 2 March 2004
online map reading, we have scanned and stored the current collection of historical maps as image segments of two different resolutions (low and high). SUCoD’s map-displaying interface can show maps arranged in multiple layers. By creating the map image datasets with the common georeferencing framework, SUCoD is able to assemble multi-layer historical maps in a chronological order for an area selected by the users. Java-enabled multi-layer interactive maps. A different type of digital map is provided on SUCoD to enable user-centred dynamic retrieval of 3D interactive city models. We have developed a Java-based map-displaying environment that can hold again multiple layers corresponding to the records of Building Age and Building Usage surveyed in the Sheffield Urban Study. Each layer of map can be turned on or off, zoomed in or out separately. Virtual reality models. 3D digital models were built manually based on the physical models. The digital models were then ‘deconstructed’ into discrete datasets that can be retrieved individually and combined to form 3D virtual city models in the virtual reality modelling language (VRML). Again, SUCoD is able to perform dynamic assembly of 3D VRML city models according to users’ selections performed on the Java-enabled maps. Related documents list. Digitised photos, drawings and textual documents were created and stored with SUCoD. We also developed a facility to enable ‘location-based’ retrieval of these urban documents. When a VRML world is reconstructed by the SUCoD server, a list of document links is also generated at the same time, pointing to documents associated with the VRML world. In addition, to enhance the usability and reusability of SUCoD’s 3D model datasets, we have developed two further interactive facilities with which users can perform online synthesis of multi-stratum virtual worlds: 앫 Upload VRML. Users can submit their own 3D designs in the VRML format to SUCoD. Any VRML model can be uploaded to SUCoD provided that the model is constructed with reference to the SUCoD’s 3D coordinate system, which can be read from SUCoD’s interactive Java maps and VRML worlds. 앫 List VRML. When interacting with SUCoD, a user can retrieve and/or upload as many VRML worlds as he or she likes. SUCoD is able to keep a track record of all the VRML worlds retrieved and/or uploaded in any single user interactive session. By activating the List VRML function, a Web page is generated in real-time, listing all the virtual
Reconstructing urban contexts
181
models retrieved or uploaded, any of which can be selected to construct a multi-stratum VRML world, resulting in a synthesis of retrieved and uploaded VRML models. A multi-stratum VRML world behaves similar to a multi-layer map, which contains multiple layers (strata) of virtual reality models. With a ‘touch sensor’ embedded within each layer, users can turn on or off any layer of VRML scene while navigating the virtual city. Given the dynamic facilities available on the SUCoD system as explained above, a user is able to reconstruct urban contexts online as displayed in maps, models and related documents. For example, by varying the spatial and temporal attributes, it is possible to reconstruct city models containing specifically ‘buildings used for light trades in the A2 B2 areas in 1900’, or ‘buildings of more than 100-year-old in the A2 B2 areas as seen in 1900’. As an example, Figure 3 shows that a hypothetical building design was uploaded to SUCoD and then combined with the site model reconstructed on SUCoD. Note that all these dynamic operations can be performed by the end-users directly over the Web without going through the ‘middle-man’ of an IT specialist.
Figure 3 Online synthesis of a hypothetical building with an urban context reconstructed on SUCoD
182
Design Studies Vol 25 No. 2 March 2004
3
Application of SUCoD in an educational setting
The development of the SUCoD system as described above originated from the need of preserving the urban study resources and making the archive materials more widely accessible through multiple digital routes. Having achieved the main objectives of data conversions and storage, we soon realised the potential of SUCoD as a platform for contemporary urban design for the city of Sheffield. We thought that if research funding permits, contemporary urban contextual datasets can be added into SUCoD just as we did for Sheffield 1900 urban data. The development of the additional ‘upload VRML’ facility is a clear demonstration that we can build a city-wide urban databank to deposit various types of historical records that will reveal how the city has evolved. The city’s contextual resources can then be used not only for a historical study purpose but also as online references to contemporary urban design. In October 2001, the SUCoD resources were first introduced to students undertaking the ‘computer applications in urban design’ course at the School of Architecture. An overall aim of the course is to provide a learning process with which students can interrelate their existing CAD skills with knowledge of urban history and experiment with SUCoD in exploring the urban history and spaces online. Working in small groups, the students were asked to achieve the following: Online urban narrative. In simple terms, urban narratives are tales (stories) about urban places or spaces. An episode of urban narrative is a story written for a specific location, which could be a scene of some urban event in the past, a description of how the urban space and/or buildings were used for an activity/usage, how an urban place is best navigated (as a tourist guide, for instance), and so on. The students organised themselves to work in groups (2–3 students per group) and produce a Web-based episode, containing the narrative composed in multimedia (e.g. 2D maps, drawings, sketches, 3D VRML models, hypertexts, etc.). Online urban design proposal. The exercise involves uploading the students’ own 3D schematic designs into SUCoD as a methodology of exploring and communicating urban design proposal online. The students were also asked to give an account of the proposal to accompany the 3D models as a Web document in a way similar to their previous creation of the urban narrative. Apart from a very basic definition, the exact interpretation of what is meant by Urban Narrative was deliberately left open. This is because that most Year 3 students at Sheffield have already gained a considerable amount of
Reconstructing urban contexts
183
Figure 4 Application SUCoD Urban
in
the
Design
of
‘Online Proposal’
project
urban experiences through their urban study trips and urban design projects located in Sheffield and in other European cities. However, some references of urban visualisation were introduced to the students, ranging from Gordon Cullen’s ‘Serial Visions’,9 Kevin Lynch’s ‘Image of a City’,10 to more casual examples such as the Dorling Kindersley Eyewitness Travel Guides. The exact format of how the urban narrative should be presented was also left to the students to decide. The only requirement was that the urban narrative must be viewable on the Web. At the beginning, the students were introduced to the SUCoD system and the digital resources made available through SUCoD over the Web. The students then decided to work on a city location (place) upon which the narrative could be developed on the basis of the SUCoD resources. The second task of proposing a schematic 3D urban design was intended for students to explore online 3D urban design visualisation using SUCoD’s ‘Upload VRML’ and ‘List VRML’ facilities. Four adjacent sites (labelled as A, B, C, D in Figure 4) as seen from the Sheffield 1900 Urban Study were designated for the online modelling project, and each student group was asked to produce a 3D design for the chosen site.
9
Cullen, G The Concise Townscape The Architectural Press, Oxford (1966) 10 Lynch, K The Image of the City The MIT Press, Cambridge, MA (1960)
184
The 3D design was not required to follow any specific urban programmes regarding what the proposed building or urban landscape was supposed to do. The primary goal of the exercise was to try out simple 3D volumes and shapes that could be ‘planted’ into the sites. Additionally, the student groups were asked to produce an urban narrative to account for the design
Design Studies Vol 25 No. 2 March 2004
development in a format similar to the previous project. To enable their digital modelling, the basic steps of preparing the VRML file, uploading and combining with site models retrieved from SUCoD were explained. The students could use any CAD packages familiar to them that would export VRML models based on their own 3D designs. To construct a multistratum VRML world containing the site and the proposed 3D design(s) successfully, the only constraints that the designers must follow are to adopt the same CAD modelling scale (1:1) and the SUCoD’s coordinate system. The 2D and 3D coordinates of any location within SUCoD can be read from the Java-enabled map and the VRML display windows. Once the modelling scale and the coordinates are established in sync with SUCoD, producing online syntheses of a site and alternative design proposals is as simple as ticking the checkboxes and pressing the buttons shown on the Web pages.
4
Unlocking digital design creativity
The Urban Narrative Online project lasted for four weeks, and eight student groups were formed. Supported by group tutorials and an interim presentation/review, eight urban narratives were produced in the end. To facilitate collaborative learning, the course tutor devised a common web site to provide links to the eight narratives submitted so that they could be viewed through a single unified interface on the Web.11 Given that the definition and purposes of the urban narrative were left open from the beginning, it is rather surprising to see that concurring strategies of narrative development have emerged among the group studies. Juxtaposition of contemporary and historical city scenes. Three groups have developed their urban narratives by piecing information together to reveal two parallel lines of illustrations: Sheffield as it is now vs. Sheffield as it was 100 years ago. For the same spatial locations or routes, the students were interested in showing the differences and similarities of the city fabric through different time frames. Virtual tourist guides. One group told their city stories simply by making short tourist guides focussing on the city routes of their choices. Scanned historical photos, captured images from the digital maps and VRML worlds are displayed as viewers navigate through various locations on the selected routes.
11
Online at http://sucod.shef .ac.uk/sucod/gallery/arc202/200 1/Proj1/project1/proj1.htm
Sequential analyses. Three groups produced what they called a ‘sequential analysis’ of the city area. Cullen’s method of serial vision was applied to establish sequences of views captured from the VRML worlds retrieved
Reconstructing urban contexts
185
from SUCoD. Textual annotations or commentaries were often added to provide further historical or contemporary information. Combined visual sequencing with juxtaposition. One group actually used a mix of juxtaposition and sequential analyses in producing their narrative study. A ‘1900—Narrative Route’ was first decided, followed by a ‘1900— Sequential Analysis’. The same 1900 route was then superimposed on a 2001 map and led to a narrative of the urban conditions in 2001. A selection of images captured from the VRML world and from the contemporary urban scene was then put side by side to form a ‘Direct Comparison’. Skills of HTML authoring were most needed in the production of the online urban narratives. Though software packages such as Microsoft FrontPage and Macromedia Dreamweaver offer handy tools in making Web pages, there are still steep learning curves for students new to these software tools. One of the most common problems students experienced was broken hyperlinks across page frames in which the authors simply got lost while trying to make connections among dozens of data files that have not been organised in a disciplined manner. More experienced students could handle the situation better and produced a compact set of HTML pages that are easy to navigate and reflect their narrative strategies (Figure 5). There
Figure 5 One group (Group 2) students’ Web authoring on the ‘Online Urban Narrative’ project
186
Design Studies Vol 25 No. 2 March 2004
seemed always a temptation for relatively inexperienced students to end up with complicated page compositions containing not properly formatted texts or graphics that actually clutter the underlying lines of narration. However, the range of HTML authoring tools offered by the software plus the online urban resources provided by SUCoD have clearly induced a fair amount of enthusiasm from the students in creating their Web presentations. Having completed the project on urban narratives, the students started to work on the online urban design proposal project for another five weeks. Apart from the locations and boundaries of the sites, the students were given very little hints as how the urban proposal might be developed. One of the key questions was that the urban sites in this case were completely virtual, that is, historical sites reconstructed electronically and exist only in virtual worlds. Again, eight design proposals were generated in the end, which were evenly distributed among the four designated sites.12 Looking at the resultant proposals, the following design approaches were observed: Contextual analyses. Using the historical maps and VRML models retrieved from SUCoD, the students were able to come up contextual analyses that influenced their proposals. Several groups came to conclude on specific building contents of their proposals by evaluating the usages of the buildings and spaces surrounding the sites. Other contextual indicators such as views, axes, accesses, street-frontage and so on also played a role. Annotated explanations produced by adding diagrams onto the maps and models were commonly seen in conveying the contextual analyses. Precedents and design references. In terms of accounting for why specific built forms were proposed, several groups pointed out explicitly the design precedents or references they have followed. Although there appeared a certain degree of arbitrariness in why particular precedents were adopted, the design references seem to be connected with how the sites were interpreted. For instance, one student perceived the site as an opportunity to place an ‘urban retreat’ and he considered the Vito Acconci’s works an inspiration, while another group pointed to Will Alsop’s Peckham Library as an example of creating a covered urban space as a transition between the proposed building and the adjacent main street (Figure 6).
12
Online at http://sucod.shef .ac.uk/sucod/gallery/arc202/200 1/Proj2/project2/proj2.htm
Shapes and volumetric modelling. In developing the urban proposal, a group of students gave themselves a list of eight ‘design issues’ ranging from height to entrance. In dealing with these issues, component shapes and volumes for the proposed building were developed. The group then took a rather innovative approach to visualising how the individual sub-
Reconstructing urban contexts
187
Figure 6 Referencing
to
other designers’ works in creating the urban design proposals.
(Below:
An
‘urban retreat’ attributed to Vito Acconci. Above: A public building following the precedent of the Peckham Library
by
the
Alsop
Architects)
volumes were shaped and interconnected by constructing a multi-stratum VRML model in which parts of the proposal were uploaded separately. Through their multi-stratum modelling, the growth of the proposed built form could therefore be explained layer by layer in relation to the surrounding urban form. In this case, the resultant multi-stratum VRML model itself becomes a compelling narrative device, as viewers could read into the developmental aspects of the urban proposal in a stepwise manner (Figure 7). Visions and colours. There are many modes of viewing that a user can take within a VRML world, such as walk, fly, study (examine) and so on with which the user can navigate to any location and look into any direction. It is therefore possible for the students to arrive at a design argument that is founded on what is meant to be seen from where. For example, one group developed a ‘glass canopy’ along their proposed building front so
188
Design Studies Vol 25 No. 2 March 2004
Figure 7 Creating a multi-stratum VRML world to show the design concepts in nine layers
that the pedestrians could still see the adjacent St. Paul’s Church while walking underneath the canopy. The contextual visions of what actually could be seen through the windows of the proposed building were also simulated through the material attribute of ‘transparency’ in VRML. The attribute of colours was another theme of urban design that the students have taken up. One group considered a ‘contrast’ could be created by introducing ‘primary colours’ into their proposed sculpture-like structure (Figure 8). Clearly, skills of 3D CAD modelling and rendering were essential in making the urban design proposals. All the VRML models submitted by the students were created mainly with 3D CAD packages that the students were familiar with. The 3D modelling capabilities of current VRML-based editors are nowhere near to some of the major CAD platforms such as AutoCAD or Microstation. Having done their 3D designs in CAD and checked the modelling scale/coordinates, the students could simply export VRML models that were combinative with sites models retrieved from SUCoD. In addition to the 3D multi-stratum VRML worlds, the students were also asked to provide commentaries on their urban proposals in a format similar to the previous project. The course tutor then set up another web site to exhibit the eight schemes in a way that each group’s VRML world and the design narrative can be viewed jointly. The provisioning of two course web sites was intended to facilitate students’ collaborative learning by browsing through fellow students’ schemes at their convenient times from wherever they may get access to the sites.
Reconstructing urban contexts
189
Figure 8 Trying out ‘glass canopy’ (above) and ‘primary colours’ (below) in relation to the urban surroundings
5
Prospects of developing an Urban Design Collaboratory
Applying digital technologies to the visualisation and simulation of the built environment at an urban scale was first seen nearly 20 years ago. However, as other researchers and educators have observed, digital systems and tools have not brought about significant benefits to urban design education and practice. This may be attributed to several practical difficulties: firstly, it is time-consuming and resource-intensive to establish an urban database covering a sufficient spatial scope; secondly, we are short of wellresearched methodologies for managing and maintaining a temporal organisation of urban data that reflect how the city may have evolved; and thirdly, the provision of innovative user-centred access facilities with which the urban designers can actively explore the urban contexts and generate design proposals is lacking. The Sheffield Urban Study programme and the creation of the SUCoD
190
Design Studies Vol 25 No. 2 March 2004
system have been pursued in an educational and academic research setting. Undoubtedly, many issues remain to be explored further. Reflecting on the Sheffield experiment, we envisage the development of an Urban Design Collaboratory supporting urban contextual study and urban design. We think that the following kinds of activities to be facilitated by the Collaboratory can be made available: (1) Creating online urban contextual resources. The Sheffield Urban Study programme has experimented with a methodology, which demonstrated how a city’s history could be studied in a systematic manner. The methodological framework developed can be applied to any geographical locations and historical periods of the city. It shows that it is possible for a School of Architecture to engage in urban study activities as a part of design education and results in an accumulative urban contextual database useful to teaching, research and design. Applying suitable digital technologies to make the urban contextual database accessible online will not only enhance its accessibility but also the extensibility of data creation over a long period of time. Central to the Collaboratory will be to achieve a sustainable long-term development of online urban contextual resources. (2) Developing dynamic urban modelling facilities. If online urban resources permit only passive browsing, the effects of supporting urban design and modelling activities will be very limited. To encourage users’ going online to perform urban design studies, online modelling applications must be developed so that users can apply to the urban resources in relation to their design interests. The dynamic retrieval and model synthesis facilities provided by SUCoD can be seen as the first step towards developing a whole range of dynamic modelling functionalities that designers can operate online directly by themselves. The Collaboratory will serve as a testing ground for experimenting with all kinds of innovative software tools relating to multidimensional and interdisciplinary urban modelling and design. For instance, in parallel with historical reconstruction, urban environmental simulation tools could be developed to investigate location-specific urban lighting, acoustics, airflow, or solar access problems. (3) Widening participation from teaching and learning. No single curriculum is able to encompass all areas of knowledge required to address the complexity encountered in today’s cities. There has been a natural fragmentation of subject areas in order to develop the necessary expertise specific to particular issues in urban design. The wide range of learning resources distributed across educational programmes makes it increasingly difficult for students to manage integration of knowledge and skills in the architectural and urban design processes. In this
Reconstructing urban contexts
191
aspect, the Urban Design Collaboratory can establish the infrastructure for sharing teaching and learning resources across various disciplinary boundaries. It can also provide a unified platform and interface for accessing online resources created in different domains. Given that the future capabilities of data exchanges are likely to grow based on the emerging technologies of Semantic Web,13 direct data communications between the Collaboratory and other software tools should be achievable. (4) Enlivening a virtual urban design forum. The traditional design review or critique processes in which students present their studies and receive criticism from their tutors or fellow students can be enhanced by the Collaboratory. Increasingly, Web-based virtual conferencing, project reviews or panellist discussions are becoming an integral part of design education (see, for instance, the Virtual Design Studio initiatives.14,15) The Urban Design Collaboratory can potentially enhance the pedagogical process a step further by providing online urban contexts as backdrop for hosting urban design forum events over the Web. Participants of these events can be collocated or geographically remote. Reviews of design development at key stages may need to be synchronous events to mark the occasions in which everybody meets at the same time. However, asynchronous and indirect virtual contacts and communications can be equally important in support of studio projects typically lasting for several months. The Collaboratory should be capable of sustaining the Virtual Urban Design Forum and facilitating educational events and activities of varying geographical and temporal conditions.
Acknowledgements The development of the SUCoD system was supported by the Arts and Humanities Research Board in the United Kingdom (RG/AN5875/APN9446). The author wishes to thank the Year 3 students at the School of Architecture, University of Sheffield, for taking part in the ARC320 course in 2001/02 session.
13
Berners-Lee, T, Hendler, J and Lassila, O ‘The semantic web’ Scientific American Vol May (2001) 29–37 14 Wojtowicz, J Virtual Design Studio Hong Kong University Press, Hong Kong (1995) 15 Shelden, D, Bharwani, S, Mitchell, W and Williams, J ‘Requirements for virtual design review’ Architectural Research Quarterly Vol 1 No winter (1995) 80–89
192
Design Studies Vol 25 No. 2 March 2004
T
he necessity of investigating and understanding an urban context in which an architectural design is to be developed has been constantly emphasised by civic authorities, design educators, critics and practitioners. At most schools of architecture, we can see that studio design projects often start with some form of contextual investigations of the city sites. The contextual study involves, typically, the gathering of information on the physical, historical and social aspects of an urban site, and the production of various kinds of records such as site drawings and models. Students then develop their proposals in relation to these contextual records. The results may turn out drastically in contrast with the existing conditions of the site or they may appear much blended into the urban setting. From a pedagogical viewpoint, one could argue that architectural design can never be entirely ‘site-determined’, yet a general consensus seems to www.elsevier.com/locate/destud 0142-694X $ - see front matter Design Studies 25 (2004) 175–192 doi:10.1016/j.destud.2003.10.005 2003 Elsevier Ltd All rights reserved Printed in Great Britain
175
1 Peng, C, Chang, D, Blundell Jones, P and Lawson, B ‘Exploring urban history and space online: design of the virtual Sheffield application’ Design Studies Vol 23 No 5 (2002) 437–453 2 Peng, C, Chang, D, Blundell Jones, P and Lawson, B ‘Making Virtual Cities Dynamic’ International Journal of Design Computing, Special Issue on Designing Virtual Worlds Vol 4, (2002) (http://www.arch.usyd.edu.au/kcdc/ journal/vol4/index.html) 3 Peng, C, Chang, D, Blundell Jones, P and Lawson, B ‘Dynamic retrievals in an urban contextual databank system using Java-CGI communications’ in B de Vries, J van Leeuwen and H Achten (eds) Proceedings of Computer Aided Architectural Design Futures 2001, 8–11 July, Eindhoven University of Technology, Eindhoven, The Netherlands, Kluwer Academic Publishers, Dordrecht (2001) pp 89–102 4 Ennis, G, Lindsay, M and Grant, M ‘VRML possibilities: the evolution of the Glasgow model’ in Proceedings of the Fifth International Conference on Virtual Systems and Multimedia (VSMM ‘99), Dundee, Scotland (1999) 5 McCullough, M and Hoinkes, R ‘Dynamic data sets as collaboration in urban design’ in M Tan and R Teh (eds) Proceedings of Computer Aided Architectural Design Futures 1995, Centre for Advanced Studies in Architecture, Department of Architecture, National University of Singapore, 24–26 September, Singapore (1995) 6 Batty, M, Dodge, M and Jiang, B ‘GIS and urban design’ in S Geertman, S Openshaw and J Stillwell (eds) Geographical Information and Planning: European Perspectives, Springer-Verlag, Berlin (1999) pp 43–65 7 Pietsch, S, Radford, A and Woodbury, R ‘Making and using a city model: Adelaide, Australia’ in Proceedings of the 19th eCAADe Conference, 29–31 August, Helsinki (2001) pp 442–447
176
subscribe to the objective of fostering the students’ appetites or inspirations for engaging in ‘dialogues’ with the urban contexts in question. Based on a recent experiment in an undergraduate urban design computing course, this paper reports a study on the usability and usefulness of a Web-based dynamic virtual city information system for directly supporting urban investigation and 3D urban design. Since late 1999, we have been developing the Sheffield Urban Contextual Databank (SUCoD) system as a Web-based open-source repository to host historical records of the city of Sheffield.1–3 We are certainly not the first to explore how digital technologies could be applied to urban studies and design. In early 1980s, the ABACUS group at the University of Strathclyde pioneered 3D computer modelling technologies for visual simulation of the city of Glasgow at a large scale, which have generated numerous educational and industrial applications.4 Working in the combined architecture and landscape domain, McCullough and Hoinkes5 considered that central to urban design are ‘dynamic, multivalent datasets for collaborative processes’, which could be better supported through a rich urban information gathering and interaction environment. Similarly, collaborating with graduate urban design students at the Oxford Brookes University, Michael Batty and co-workers6] investigated how geographical information systems (GIS) could be linked with computer-aided design (CAD) to provide intuitive GIS-driven visualisation and communication facilities for participative urban design. More recently, Susan Pietsch and colleagues7 reported their studies of using sections of the city models built for the city of Adelaide in student design projects. As seen from the related studies, the creation of urban contextual resources at a large scale and applying the digital resources in architectural and urban design still present considerable theoretical and technical challenges.
1
Studying the historical city fabric of Sheffield
The ‘Sheffield Urban Study’ programme was first launched in 1998 at the School of Architecture, University of Sheffield. The starting point for the Urban Study was a consensus among the staff that context is important, and that one should always know the history of a site in order to understand how it was formed and what it means, whatever use is eventually made of that information. By accessing and studying the historical maps and records from the City Archives, Local Studies Centre and the Hawley Collection, students at the Diploma School during 1998–2001 built a physical model of the city as it stood in the year 1900, at a scale of 1:500. This was accomplished by dividing the city with a grid into 200 m2, each of which was investigated and modelled by a group of four or five students. They were asked to report on the ages and uses of all the buildings at the
Design Studies Vol 25 No. 2 March 2004
Figure 1 A north-south aerial view of the assembled Sheffield 1900 model
period under study. Each student was also asked to choose and study a particular building, comparing it with the equivalent building today if appropriate.8 To date, a total of 41 unit models have been constructed, revealing a view of historic Sheffield unseen before (Figure 1). Since its inception, the urban study archive has grown progressively following the urban grid as an analytical framework which can be extended at anytime to cover other areas of the city. To the research community, we believe that the Sheffield Urban Study resources have the following potential values:
8 Blundell Jones, P, Williams, A and Lintonbon, J ‘The Sheffield urban study project’ Architectural Research Quarterly Vol 3 No 3 (1999) 235–244
앫 Heritage. The city of Sheffield grew from little more than a village to one of the largest cities in Britain between 1750 and 1900 because of its industries and the world-famous Sheffield steel. The heavy steel industry continues in mechanised form far outside the city, but the tool and cutlery industries have almost faded away through international competition. Though initiated by it, the city has been slow to appreciate the value of its industrial heritage. The industrial installations and buildings, usually in very poor condition at the point of economic collapse, have very largely been razed to the ground. Even the street pattern has in many cases been abandoned. Yet from a heritage point of view, Shef-
Reconstructing urban contexts
177
앫
앫
앫
앫
field could be of great international interest, and could do much more to celebrate its origins and industrial heroes like Benjamin Huntsman, inventor of the crucible process. Every year, however, more industrial remains are destroyed. The Sheffield Urban Study programme systematically identifies the history of every site, giving a necessary overview of the layers of the city and their relative importance. Current planning and understanding of context. Developers and architects tend to regard a site as a blank slate unless buildings on it are listed, yet often it is possible to preserve some characteristics or even parts of existing buildings within a new design to give it some historical continuity. With a database such as can be provided through the Urban Study, the evidence is at least quickly to hand and the possibility can be considered. Student experience. As an educational course for architectural students, the Urban Study engenders a personal awareness of quality of place and of the layers left by history, so never again can a site be regarded as ‘a nowhere place’. Involvement in compiling the database teaches students techniques of recording and analysis, the use of sources such as archives, and the kind of thing to be sought and found in a typical urban site. It is a necessary training for those whose responsibilities include some duty to preserve as well as to change the nature of places. Database. The Urban Study produces a unique record of the city, which helps in understanding its form and social history. With an increase in resources, this could be made available to a much wider circle of users. A generic methodology. The Urban Study has pioneered new techniques in historical analysis and recording, both in its initial paper-based and physical modelling phase, and through its translation into an electronic version. In essence, the methodology can be re-applied to any other city, and not just industrial ones. Although it might initially appear straightforward, the translation into electronic form has revealed many difficulties and required considerable reflection about methods of storing and classifying information, and about the kinds of information that lend themselves to electronic media. In some areas, they have great advantages over paper and in others they are lacking. The greatest challenge has been to create classification methods that remain to some extent open in order to embrace types of data as yet unforeseen.
2
The Sheffield Urban Contextual Databank (SUCoD) system
As it is, the Sheffield 1900 database has taken the form of a paper file and a cardboard model of each grid square. The ‘box-file’ of documents for every grid square has proved a most practical way of assembling data
178
Design Studies Vol 25 No. 2 March 2004
initially. However, we find that uses of the physical state of the data are limited in several aspects: (a) The contextual data acquired in its current formats do not allow for efficient storage, transportation, management, and rapid reproduction. Clearly, the physical state of the data does not facilitate more widely collaborative research and debate on proposing new architectural interventions in an educational or professional setting. Also, the graphic conventions and formats used are not altogether consistent. (b)Urban contextual study is potentially a continuous process undertaken by academic or professional design research groups, and material for the database may be accumulated over years. A physical approach to modelling and documentation has limited flexibility for extending a database consistently over a long period. (c) A physical approach can follow a clear organisation if parts of the contextual study are geographically tightly connected as with the Sheffield 1900 project, but this is not always the case. Contextual inquiries may involve a region with the parts geographically dispersed, and with no imminent connections to be made. In such cases, it is difficult to establish the organisation through direct physical modelling. (d)The aim of understanding and showing how places have evolved over a period of time can be achieved through juxtaposition of a series of maps, drawings or models. To display the evolutions three-dimensionally in physical models, however, is cumbersome if not impossible. Simply replacing parts of the model does not offer optical continuity or visual cues for appreciating processes as they change.
Parallel to the Sheffield Urban Study project, we have been investigating how to organise the urban contextual information electronically, taking into account the purposes and requirements of the study as already discussed. Our aim is to establish an effective methodology for creating and managing multidimensional urban datasets that could support our design research and education programmes. We also take into account the unique requirements of Sheffield as a post-industrial city and the potential usefulness of research outcomes. Over the past two years, we have designed and implemented the Sheffield Urban Contextual Databank (SUCoD) system as a pilot study of digitising the urban resources. SUCoD is able to deliver multidimensional urban datasets such as interactive maps and virtual reality models by responding to users’ choices and requests transmitted over the Web
Reconstructing urban contexts
179
(http://sucod.shef.ac.uk). In developing the system, we were particularly interested in experimenting with an alternative methodology that will lift some of the limitations seen in the existing approaches to building ‘virtual cities’, which have been largely based on static urban modelling. In our view, existing methods of making 3D digital city models pose fundamental difficulties in making these models applicable to directly supporting contextual urban study, planning, and design. We commonly see that the basic recognition and definition of an urban context often varies from one individual to another. An architect, for instance, may take the entire city or town as the contextual framework to work with when designing a primary school for a certain location, while another architect may choose to focus on a specific cluster of buildings surrounding the site as the sole context to be addressed. We believe that it is essential for an urban information system to allow the exercise of individual differences in recognising and defining the focal locations and boundaries of an urban context, as there lies the origin of creativity that may contribute to qualities of urban design in a fundamental way. In its current release, SUCoD provides users with map-based interface to select and retrieve the following kinds of urban contextual resources (Figure 2). Scanned historical maps. Historical maps are of an important type of urban contextual resource that users may read with a great interest. To enable
Figure 2 The (SUCoD) system on the Web as seen through Netscape
180
Design Studies Vol 25 No. 2 March 2004
online map reading, we have scanned and stored the current collection of historical maps as image segments of two different resolutions (low and high). SUCoD’s map-displaying interface can show maps arranged in multiple layers. By creating the map image datasets with the common georeferencing framework, SUCoD is able to assemble multi-layer historical maps in a chronological order for an area selected by the users. Java-enabled multi-layer interactive maps. A different type of digital map is provided on SUCoD to enable user-centred dynamic retrieval of 3D interactive city models. We have developed a Java-based map-displaying environment that can hold again multiple layers corresponding to the records of Building Age and Building Usage surveyed in the Sheffield Urban Study. Each layer of map can be turned on or off, zoomed in or out separately. Virtual reality models. 3D digital models were built manually based on the physical models. The digital models were then ‘deconstructed’ into discrete datasets that can be retrieved individually and combined to form 3D virtual city models in the virtual reality modelling language (VRML). Again, SUCoD is able to perform dynamic assembly of 3D VRML city models according to users’ selections performed on the Java-enabled maps. Related documents list. Digitised photos, drawings and textual documents were created and stored with SUCoD. We also developed a facility to enable ‘location-based’ retrieval of these urban documents. When a VRML world is reconstructed by the SUCoD server, a list of document links is also generated at the same time, pointing to documents associated with the VRML world. In addition, to enhance the usability and reusability of SUCoD’s 3D model datasets, we have developed two further interactive facilities with which users can perform online synthesis of multi-stratum virtual worlds: 앫 Upload VRML. Users can submit their own 3D designs in the VRML format to SUCoD. Any VRML model can be uploaded to SUCoD provided that the model is constructed with reference to the SUCoD’s 3D coordinate system, which can be read from SUCoD’s interactive Java maps and VRML worlds. 앫 List VRML. When interacting with SUCoD, a user can retrieve and/or upload as many VRML worlds as he or she likes. SUCoD is able to keep a track record of all the VRML worlds retrieved and/or uploaded in any single user interactive session. By activating the List VRML function, a Web page is generated in real-time, listing all the virtual
Reconstructing urban contexts
181
models retrieved or uploaded, any of which can be selected to construct a multi-stratum VRML world, resulting in a synthesis of retrieved and uploaded VRML models. A multi-stratum VRML world behaves similar to a multi-layer map, which contains multiple layers (strata) of virtual reality models. With a ‘touch sensor’ embedded within each layer, users can turn on or off any layer of VRML scene while navigating the virtual city. Given the dynamic facilities available on the SUCoD system as explained above, a user is able to reconstruct urban contexts online as displayed in maps, models and related documents. For example, by varying the spatial and temporal attributes, it is possible to reconstruct city models containing specifically ‘buildings used for light trades in the A2 B2 areas in 1900’, or ‘buildings of more than 100-year-old in the A2 B2 areas as seen in 1900’. As an example, Figure 3 shows that a hypothetical building design was uploaded to SUCoD and then combined with the site model reconstructed on SUCoD. Note that all these dynamic operations can be performed by the end-users directly over the Web without going through the ‘middle-man’ of an IT specialist.
Figure 3 Online synthesis of a hypothetical building with an urban context reconstructed on SUCoD
182
Design Studies Vol 25 No. 2 March 2004
3
Application of SUCoD in an educational setting
The development of the SUCoD system as described above originated from the need of preserving the urban study resources and making the archive materials more widely accessible through multiple digital routes. Having achieved the main objectives of data conversions and storage, we soon realised the potential of SUCoD as a platform for contemporary urban design for the city of Sheffield. We thought that if research funding permits, contemporary urban contextual datasets can be added into SUCoD just as we did for Sheffield 1900 urban data. The development of the additional ‘upload VRML’ facility is a clear demonstration that we can build a city-wide urban databank to deposit various types of historical records that will reveal how the city has evolved. The city’s contextual resources can then be used not only for a historical study purpose but also as online references to contemporary urban design. In October 2001, the SUCoD resources were first introduced to students undertaking the ‘computer applications in urban design’ course at the School of Architecture. An overall aim of the course is to provide a learning process with which students can interrelate their existing CAD skills with knowledge of urban history and experiment with SUCoD in exploring the urban history and spaces online. Working in small groups, the students were asked to achieve the following: Online urban narrative. In simple terms, urban narratives are tales (stories) about urban places or spaces. An episode of urban narrative is a story written for a specific location, which could be a scene of some urban event in the past, a description of how the urban space and/or buildings were used for an activity/usage, how an urban place is best navigated (as a tourist guide, for instance), and so on. The students organised themselves to work in groups (2–3 students per group) and produce a Web-based episode, containing the narrative composed in multimedia (e.g. 2D maps, drawings, sketches, 3D VRML models, hypertexts, etc.). Online urban design proposal. The exercise involves uploading the students’ own 3D schematic designs into SUCoD as a methodology of exploring and communicating urban design proposal online. The students were also asked to give an account of the proposal to accompany the 3D models as a Web document in a way similar to their previous creation of the urban narrative. Apart from a very basic definition, the exact interpretation of what is meant by Urban Narrative was deliberately left open. This is because that most Year 3 students at Sheffield have already gained a considerable amount of
Reconstructing urban contexts
183
Figure 4 Application SUCoD Urban
in
the
Design
of
‘Online Proposal’
project
urban experiences through their urban study trips and urban design projects located in Sheffield and in other European cities. However, some references of urban visualisation were introduced to the students, ranging from Gordon Cullen’s ‘Serial Visions’,9 Kevin Lynch’s ‘Image of a City’,10 to more casual examples such as the Dorling Kindersley Eyewitness Travel Guides. The exact format of how the urban narrative should be presented was also left to the students to decide. The only requirement was that the urban narrative must be viewable on the Web. At the beginning, the students were introduced to the SUCoD system and the digital resources made available through SUCoD over the Web. The students then decided to work on a city location (place) upon which the narrative could be developed on the basis of the SUCoD resources. The second task of proposing a schematic 3D urban design was intended for students to explore online 3D urban design visualisation using SUCoD’s ‘Upload VRML’ and ‘List VRML’ facilities. Four adjacent sites (labelled as A, B, C, D in Figure 4) as seen from the Sheffield 1900 Urban Study were designated for the online modelling project, and each student group was asked to produce a 3D design for the chosen site.
9
Cullen, G The Concise Townscape The Architectural Press, Oxford (1966) 10 Lynch, K The Image of the City The MIT Press, Cambridge, MA (1960)
184
The 3D design was not required to follow any specific urban programmes regarding what the proposed building or urban landscape was supposed to do. The primary goal of the exercise was to try out simple 3D volumes and shapes that could be ‘planted’ into the sites. Additionally, the student groups were asked to produce an urban narrative to account for the design
Design Studies Vol 25 No. 2 March 2004
development in a format similar to the previous project. To enable their digital modelling, the basic steps of preparing the VRML file, uploading and combining with site models retrieved from SUCoD were explained. The students could use any CAD packages familiar to them that would export VRML models based on their own 3D designs. To construct a multistratum VRML world containing the site and the proposed 3D design(s) successfully, the only constraints that the designers must follow are to adopt the same CAD modelling scale (1:1) and the SUCoD’s coordinate system. The 2D and 3D coordinates of any location within SUCoD can be read from the Java-enabled map and the VRML display windows. Once the modelling scale and the coordinates are established in sync with SUCoD, producing online syntheses of a site and alternative design proposals is as simple as ticking the checkboxes and pressing the buttons shown on the Web pages.
4
Unlocking digital design creativity
The Urban Narrative Online project lasted for four weeks, and eight student groups were formed. Supported by group tutorials and an interim presentation/review, eight urban narratives were produced in the end. To facilitate collaborative learning, the course tutor devised a common web site to provide links to the eight narratives submitted so that they could be viewed through a single unified interface on the Web.11 Given that the definition and purposes of the urban narrative were left open from the beginning, it is rather surprising to see that concurring strategies of narrative development have emerged among the group studies. Juxtaposition of contemporary and historical city scenes. Three groups have developed their urban narratives by piecing information together to reveal two parallel lines of illustrations: Sheffield as it is now vs. Sheffield as it was 100 years ago. For the same spatial locations or routes, the students were interested in showing the differences and similarities of the city fabric through different time frames. Virtual tourist guides. One group told their city stories simply by making short tourist guides focussing on the city routes of their choices. Scanned historical photos, captured images from the digital maps and VRML worlds are displayed as viewers navigate through various locations on the selected routes.
11
Online at http://sucod.shef .ac.uk/sucod/gallery/arc202/200 1/Proj1/project1/proj1.htm
Sequential analyses. Three groups produced what they called a ‘sequential analysis’ of the city area. Cullen’s method of serial vision was applied to establish sequences of views captured from the VRML worlds retrieved
Reconstructing urban contexts
185
from SUCoD. Textual annotations or commentaries were often added to provide further historical or contemporary information. Combined visual sequencing with juxtaposition. One group actually used a mix of juxtaposition and sequential analyses in producing their narrative study. A ‘1900—Narrative Route’ was first decided, followed by a ‘1900— Sequential Analysis’. The same 1900 route was then superimposed on a 2001 map and led to a narrative of the urban conditions in 2001. A selection of images captured from the VRML world and from the contemporary urban scene was then put side by side to form a ‘Direct Comparison’. Skills of HTML authoring were most needed in the production of the online urban narratives. Though software packages such as Microsoft FrontPage and Macromedia Dreamweaver offer handy tools in making Web pages, there are still steep learning curves for students new to these software tools. One of the most common problems students experienced was broken hyperlinks across page frames in which the authors simply got lost while trying to make connections among dozens of data files that have not been organised in a disciplined manner. More experienced students could handle the situation better and produced a compact set of HTML pages that are easy to navigate and reflect their narrative strategies (Figure 5). There
Figure 5 One group (Group 2) students’ Web authoring on the ‘Online Urban Narrative’ project
186
Design Studies Vol 25 No. 2 March 2004
seemed always a temptation for relatively inexperienced students to end up with complicated page compositions containing not properly formatted texts or graphics that actually clutter the underlying lines of narration. However, the range of HTML authoring tools offered by the software plus the online urban resources provided by SUCoD have clearly induced a fair amount of enthusiasm from the students in creating their Web presentations. Having completed the project on urban narratives, the students started to work on the online urban design proposal project for another five weeks. Apart from the locations and boundaries of the sites, the students were given very little hints as how the urban proposal might be developed. One of the key questions was that the urban sites in this case were completely virtual, that is, historical sites reconstructed electronically and exist only in virtual worlds. Again, eight design proposals were generated in the end, which were evenly distributed among the four designated sites.12 Looking at the resultant proposals, the following design approaches were observed: Contextual analyses. Using the historical maps and VRML models retrieved from SUCoD, the students were able to come up contextual analyses that influenced their proposals. Several groups came to conclude on specific building contents of their proposals by evaluating the usages of the buildings and spaces surrounding the sites. Other contextual indicators such as views, axes, accesses, street-frontage and so on also played a role. Annotated explanations produced by adding diagrams onto the maps and models were commonly seen in conveying the contextual analyses. Precedents and design references. In terms of accounting for why specific built forms were proposed, several groups pointed out explicitly the design precedents or references they have followed. Although there appeared a certain degree of arbitrariness in why particular precedents were adopted, the design references seem to be connected with how the sites were interpreted. For instance, one student perceived the site as an opportunity to place an ‘urban retreat’ and he considered the Vito Acconci’s works an inspiration, while another group pointed to Will Alsop’s Peckham Library as an example of creating a covered urban space as a transition between the proposed building and the adjacent main street (Figure 6).
12
Online at http://sucod.shef .ac.uk/sucod/gallery/arc202/200 1/Proj2/project2/proj2.htm
Shapes and volumetric modelling. In developing the urban proposal, a group of students gave themselves a list of eight ‘design issues’ ranging from height to entrance. In dealing with these issues, component shapes and volumes for the proposed building were developed. The group then took a rather innovative approach to visualising how the individual sub-
Reconstructing urban contexts
187
Figure 6 Referencing
to
other designers’ works in creating the urban design proposals.
(Below:
An
‘urban retreat’ attributed to Vito Acconci. Above: A public building following the precedent of the Peckham Library
by
the
Alsop
Architects)
volumes were shaped and interconnected by constructing a multi-stratum VRML model in which parts of the proposal were uploaded separately. Through their multi-stratum modelling, the growth of the proposed built form could therefore be explained layer by layer in relation to the surrounding urban form. In this case, the resultant multi-stratum VRML model itself becomes a compelling narrative device, as viewers could read into the developmental aspects of the urban proposal in a stepwise manner (Figure 7). Visions and colours. There are many modes of viewing that a user can take within a VRML world, such as walk, fly, study (examine) and so on with which the user can navigate to any location and look into any direction. It is therefore possible for the students to arrive at a design argument that is founded on what is meant to be seen from where. For example, one group developed a ‘glass canopy’ along their proposed building front so
188
Design Studies Vol 25 No. 2 March 2004
Figure 7 Creating a multi-stratum VRML world to show the design concepts in nine layers
that the pedestrians could still see the adjacent St. Paul’s Church while walking underneath the canopy. The contextual visions of what actually could be seen through the windows of the proposed building were also simulated through the material attribute of ‘transparency’ in VRML. The attribute of colours was another theme of urban design that the students have taken up. One group considered a ‘contrast’ could be created by introducing ‘primary colours’ into their proposed sculpture-like structure (Figure 8). Clearly, skills of 3D CAD modelling and rendering were essential in making the urban design proposals. All the VRML models submitted by the students were created mainly with 3D CAD packages that the students were familiar with. The 3D modelling capabilities of current VRML-based editors are nowhere near to some of the major CAD platforms such as AutoCAD or Microstation. Having done their 3D designs in CAD and checked the modelling scale/coordinates, the students could simply export VRML models that were combinative with sites models retrieved from SUCoD. In addition to the 3D multi-stratum VRML worlds, the students were also asked to provide commentaries on their urban proposals in a format similar to the previous project. The course tutor then set up another web site to exhibit the eight schemes in a way that each group’s VRML world and the design narrative can be viewed jointly. The provisioning of two course web sites was intended to facilitate students’ collaborative learning by browsing through fellow students’ schemes at their convenient times from wherever they may get access to the sites.
Reconstructing urban contexts
189
Figure 8 Trying out ‘glass canopy’ (above) and ‘primary colours’ (below) in relation to the urban surroundings
5
Prospects of developing an Urban Design Collaboratory
Applying digital technologies to the visualisation and simulation of the built environment at an urban scale was first seen nearly 20 years ago. However, as other researchers and educators have observed, digital systems and tools have not brought about significant benefits to urban design education and practice. This may be attributed to several practical difficulties: firstly, it is time-consuming and resource-intensive to establish an urban database covering a sufficient spatial scope; secondly, we are short of wellresearched methodologies for managing and maintaining a temporal organisation of urban data that reflect how the city may have evolved; and thirdly, the provision of innovative user-centred access facilities with which the urban designers can actively explore the urban contexts and generate design proposals is lacking. The Sheffield Urban Study programme and the creation of the SUCoD
190
Design Studies Vol 25 No. 2 March 2004
system have been pursued in an educational and academic research setting. Undoubtedly, many issues remain to be explored further. Reflecting on the Sheffield experiment, we envisage the development of an Urban Design Collaboratory supporting urban contextual study and urban design. We think that the following kinds of activities to be facilitated by the Collaboratory can be made available: (1) Creating online urban contextual resources. The Sheffield Urban Study programme has experimented with a methodology, which demonstrated how a city’s history could be studied in a systematic manner. The methodological framework developed can be applied to any geographical locations and historical periods of the city. It shows that it is possible for a School of Architecture to engage in urban study activities as a part of design education and results in an accumulative urban contextual database useful to teaching, research and design. Applying suitable digital technologies to make the urban contextual database accessible online will not only enhance its accessibility but also the extensibility of data creation over a long period of time. Central to the Collaboratory will be to achieve a sustainable long-term development of online urban contextual resources. (2) Developing dynamic urban modelling facilities. If online urban resources permit only passive browsing, the effects of supporting urban design and modelling activities will be very limited. To encourage users’ going online to perform urban design studies, online modelling applications must be developed so that users can apply to the urban resources in relation to their design interests. The dynamic retrieval and model synthesis facilities provided by SUCoD can be seen as the first step towards developing a whole range of dynamic modelling functionalities that designers can operate online directly by themselves. The Collaboratory will serve as a testing ground for experimenting with all kinds of innovative software tools relating to multidimensional and interdisciplinary urban modelling and design. For instance, in parallel with historical reconstruction, urban environmental simulation tools could be developed to investigate location-specific urban lighting, acoustics, airflow, or solar access problems. (3) Widening participation from teaching and learning. No single curriculum is able to encompass all areas of knowledge required to address the complexity encountered in today’s cities. There has been a natural fragmentation of subject areas in order to develop the necessary expertise specific to particular issues in urban design. The wide range of learning resources distributed across educational programmes makes it increasingly difficult for students to manage integration of knowledge and skills in the architectural and urban design processes. In this
Reconstructing urban contexts
191
aspect, the Urban Design Collaboratory can establish the infrastructure for sharing teaching and learning resources across various disciplinary boundaries. It can also provide a unified platform and interface for accessing online resources created in different domains. Given that the future capabilities of data exchanges are likely to grow based on the emerging technologies of Semantic Web,13 direct data communications between the Collaboratory and other software tools should be achievable. (4) Enlivening a virtual urban design forum. The traditional design review or critique processes in which students present their studies and receive criticism from their tutors or fellow students can be enhanced by the Collaboratory. Increasingly, Web-based virtual conferencing, project reviews or panellist discussions are becoming an integral part of design education (see, for instance, the Virtual Design Studio initiatives.14,15) The Urban Design Collaboratory can potentially enhance the pedagogical process a step further by providing online urban contexts as backdrop for hosting urban design forum events over the Web. Participants of these events can be collocated or geographically remote. Reviews of design development at key stages may need to be synchronous events to mark the occasions in which everybody meets at the same time. However, asynchronous and indirect virtual contacts and communications can be equally important in support of studio projects typically lasting for several months. The Collaboratory should be capable of sustaining the Virtual Urban Design Forum and facilitating educational events and activities of varying geographical and temporal conditions.
Acknowledgements The development of the SUCoD system was supported by the Arts and Humanities Research Board in the United Kingdom (RG/AN5875/APN9446). The author wishes to thank the Year 3 students at the School of Architecture, University of Sheffield, for taking part in the ARC320 course in 2001/02 session.
13
Berners-Lee, T, Hendler, J and Lassila, O ‘The semantic web’ Scientific American Vol May (2001) 29–37 14 Wojtowicz, J Virtual Design Studio Hong Kong University Press, Hong Kong (1995) 15 Shelden, D, Bharwani, S, Mitchell, W and Williams, J ‘Requirements for virtual design review’ Architectural Research Quarterly Vol 1 No winter (1995) 80–89
192
Design Studies Vol 25 No. 2 March 2004