10
Comparison of CAD and manual sketching tools for teaching architectural design Rahinah Ibrahim a, , Farzad Pour Rahimian b a Department of Architecture, Faculty of Design and Architecture, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor D.E., Malaysia b School of Built and Natural Environment, University of Central Lancashire, UK abstract article info Keywords: Conceptual design VR 3D sketching Design collaboration CAD Protocol analysis Communications for information synchronisation during the conceptual design phase usually require designers to employ more exible and intuitive digital design tools. In developing such support tools, a case study research was initiated to rst understand the current state of communication among novice design team members. The main purpose of the study is to understand the novice designers' collaboration culture when working on conceptual architectural design projects. The overall case study research involves 1) ethnography for data collection and 2) artefact and protocol analyses for data analysis. This paper presents the latter data analyses based on the earlier ethnography results about novice designers' conceptual design and technology design synthesis strategies. Data analyses on the characteristics of utilised external representation tools indicate that although conventional manual sketching is benecial for providing rich intuitive design concepts, they have limitations when novice designers need to oversee complicated design problems. The study also found that although current conventional CAD tools are advantageous for detailed engineering design articulation, they do hinder novice designers' creativity due to their intuitive ideation limitation. Based on these results, this paper discusses the identied advantages and challenges of current design media and then proposes an alternative VR-based design interface for enhancing cognition and communication among designers during the conceptual design phase. © 2010 Elsevier B.V. All rights reserved. 1. Introduction With progressive globalisation and specialisation trends in the building industry, collaboration among design stakeholders in distant locations becomes crucial [40,45]. Computer Supported Collaborative Works (CSCWs) [45] are currently no longer mere facilities but an integral part of comprehensive architecture, engineering and con- struction (AEC) rms. An earlier study [32] proposes the use of high- tech visualisation media for digitisation of the conceptual architec- tural design process. Furthermore, some studies (e.g. [16] and [19]) recommend integration of design and construction process to support collaboration among team members. Moreover, they argue that integration has major advantages in decreasing labour and material costs within current comprehensive production procedure models. This integration is yet to happen smoothly since heterogeneous design media are being used during different design stages. For in- stance, conventional analogue format of design ideation tools, which are used during early conceptual design phase, are yet to be replaced with appropriate digital formats. This replacement is much desired for helping conceptual design stage outputs to better t into the suc- ceeding computerised engineering stages of the design process. We posit that the disintegration of the design parts is due to the limited efciency of conventional Computer Aided Design (CAD) software during the intuitive conceptual design activities [25], causing de- signers to struggle in ensuring the transfer of information from the conceptual architectural design to rational engineering parts. Another reason is that the majority of existing geometric modelling software entail a high degree of specialisation from the users in order to achieve the nal forms that designers desire. However, not all designers can and need to reach this distinctive degree of skills [27]. The problem is magnied when novice designers have limited expertise in using an external representation tool as a means for expressing and fortifying their design ideas. Consequently, such constricted approach hampers the capability of the design process and the collaboration that goes along [25] with it. Therefore, it can cause miscommunication for novice designers. An example is the case of architectural education where students and studio masters desire clear communication means during design studio tutoring. It is our intent to use this study to formulate a theoretical foun- dation for developing an alternative digital design medium for better supporting the conceptual architectural design education. Our main aim is to understand the current state of communication among novice design team members. Therefore, we conducted a case study research to understand novice designers' collaboration culture when Automation in Construction 19 (2010) 978987 Corresponding author. E-mail addresses: [email protected] (R. Ibrahim), [email protected] (F. Pour Rahimian). 0926-5805/$ see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.autcon.2010.09.003 Contents lists available at ScienceDirect Automation in Construction journal homepage: www.elsevier.com/locate/autcon

Comparison of CAD and manual sketching tools for teaching architectural design

Embed Size (px)

Citation preview

Automation in Construction 19 (2010) 978–987

Contents lists available at ScienceDirect

Automation in Construction

j ourna l homepage: www.e lsev ie r.com/ locate /autcon

Comparison of CAD and manual sketching tools for teaching architectural design

Rahinah Ibrahim a,⁎, Farzad Pour Rahimian b

a Department of Architecture, Faculty of Design and Architecture, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor D.E., Malaysiab School of Built and Natural Environment, University of Central Lancashire, UK

⁎ Corresponding author.E-mail addresses: [email protected] (R. Ib

[email protected] (F. Pour Rahimian).

0926-5805/$ – see front matter © 2010 Elsevier B.V. Adoi:10.1016/j.autcon.2010.09.003

a b s t r a c t

a r t i c l e i n f o

Keywords:

Conceptual designVR 3D sketchingDesign collaborationCADProtocol analysis

Communications for information synchronisation during the conceptual design phase usually requiredesigners to employ more flexible and intuitive digital design tools. In developing such support tools, a casestudy research was initiated to first understand the current state of communication among novice designteam members. The main purpose of the study is to understand the novice designers' collaboration culturewhen working on conceptual architectural design projects. The overall case study research involves 1)ethnography for data collection and 2) artefact and protocol analyses for data analysis. This paper presents thelatter data analyses based on the earlier ethnography results about novice designers' conceptual design andtechnology design synthesis strategies. Data analyses on the characteristics of utilised external representationtools indicate that although conventional manual sketching is beneficial for providing rich intuitive designconcepts, they have limitations when novice designers need to oversee complicated design problems. Thestudy also found that although current conventional CAD tools are advantageous for detailed engineeringdesign articulation, they do hinder novice designers' creativity due to their intuitive ideation limitation. Basedon these results, this paper discusses the identified advantages and challenges of current design media andthen proposes an alternative VR-based design interface for enhancing cognition and communication amongdesigners during the conceptual design phase.

rahim),

ll rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

With progressive globalisation and specialisation trends in thebuilding industry, collaboration among design stakeholders in distantlocations becomes crucial [40,45]. Computer Supported CollaborativeWorks (CSCWs) [45] are currently no longer mere facilities but anintegral part of comprehensive architecture, engineering and con-struction (AEC) firms. An earlier study [32] proposes the use of high-tech visualisation media for digitisation of the conceptual architec-tural design process. Furthermore, some studies (e.g. [16] and [19])recommend integration of design and construction process to supportcollaboration among team members. Moreover, they argue thatintegration has major advantages in decreasing labour and materialcosts within current comprehensive production procedure models.

This integration is yet to happen smoothly since heterogeneousdesign media are being used during different design stages. For in-stance, conventional analogue format of design ideation tools, whichare used during early conceptual design phase, are yet to be replacedwith appropriate digital formats. This replacement is much desired forhelping conceptual design stage outputs to better fit into the suc-

ceeding computerised engineering stages of the design process. Weposit that the disintegration of the design parts is due to the limitedefficiency of conventional Computer Aided Design (CAD) softwareduring the intuitive conceptual design activities [25], causing de-signers to struggle in ensuring the transfer of information from theconceptual architectural design to rational engineering parts. Anotherreason is that the majority of existing geometric modelling softwareentail a high degree of specialisation from the users in order to achievethe final forms that designers desire. However, not all designers canand need to reach this distinctive degree of skills [27]. The problem ismagnified when novice designers have limited expertise in using anexternal representation tool as a means for expressing and fortifyingtheir design ideas. Consequently, such constricted approach hampersthe capability of the design process and the collaboration that goesalong [25] with it. Therefore, it can cause miscommunication fornovice designers. An example is the case of architectural educationwhere students and studio masters desire clear communicationmeans during design studio tutoring.

It is our intent to use this study to formulate a theoretical foun-dation for developing an alternative digital design medium for bettersupporting the conceptual architectural design education. Our mainaim is to understand the current state of communication amongnovice design team members. Therefore, we conducted a case studyresearch to understand novice designers' collaboration culture when

979R. Ibrahim, F. Pour Rahimian / Automation in Construction 19 (2010) 978–987

working on conceptual architectural design projects. The overall casestudy research involves 1) ethnography for data collection and 2)artefact and protocol analyses for data analysis. This paper presentsthe latter empirical data analyses based on the earlier ethnographyresults about conceptual design and technology design strategiesduring the conceptual architectural design phase.

The purpose of the conducted artefact and protocol analyses wasto objectively compare the design synthesis' characteristics whenarchitectural students utilise external representation tools during theconceptual architectural design phase. In this paper, we summarisethe earlier ethnography findings about conceptual design and tech-nology strategies used by novice designers during conceptual archi-tectural design, explain the artefact and protocol analyses part of thecase study research methodology and present the resulting char-acteristics of utilised current external representation tools duringcollaborative design process. We conclude the paper by discussing theidentified advantages and challenges of the current design mediabefore proposing recommendations towards the development of aVR-based alternative design interface that would improve designrepresentation, hence, enhancing cognition and communicationamong novice designers during the conceptual design phase.

2. Literature review

This paper presents a literature survey on two areas of design,which are design representations, and challenges and benefits ofthe current design tools. We focused on literature for teaching andlearning design since design methodologies are usually trans-ferred to novice designers by professional practitioners in designstudios.

Many studies generally acknowledge design as a learning do-main, in which past experiences and basic techniques are bigcontributors. However, most argue that the obtained skill is rarelyexplicit that even experienced designers have difficulties in articu-lating and understanding their own design proficiencies [42]. Ourstudy has noted the diverse research agendas in design thinking.These include internal and external representations of designedobjects [1], issues on design generation [8], knowledge bases ofdesign thinking [44], formulation of design problems [2] and thoughtprocesses applied to learning [37].

Despite these multiple varied studies, some studies (e.g. [39] and[42]) continue recommending the need for a design methodology forimproving the design process and the quality of designed artefacts.Simultaneously, the early stages of problem finding and analysisduring semantic design phase are of central importance to designresults because 70% to 80% of production overheads are determinedthere [39]. Moreover, the characteristics of these stages are stronglyaffected by the tool that designers use when embodying theirdesigned concepts [33]. Scholars (e.g. [6] and [43]) have suggestedthat among the design tools, external representations not only serveas memory aids, but also facilitate or constrain the inference, theproblem-solving and the design stage understanding during ideageneration process. Other researchers (e.g. [39] and [42]) believe thatdesigners can benefit from abstract external representations (i.e.sketches) particularly when they are in the early conceptual designstages. However, despite the diversity of available external represen-tation tools, we believe that there is a need for a design representationtool that could support conceptual design collaboration whendesigners need external representations the most.

Existing literature highlights sketching using pencils and papersas one of the most famous abstract representation methods. By theyear 2000, its effectiveness—particularly during early conceptualdesign stages—was frequently appreciated (e.g. [6], [14], [17], [26],and [35]). This advocacy was at the highest level when somescholars (e.g. [38], and [4]) highlighted the importance of freehandsketches as an indispensable media for designers to make reflective

dialogues with their own ideas. Such gratitude to manual sketchingmethodologies started waning with improvements of ComputerAided Design (CAD) tools and their increasing utilisation in complexprojects due to globalisation challenges. A more important reasonfor this tendency towards digitisation of a design process is theadded value of digital representation for future analysis and processintegration.

Our literature survey notes the increasing tendency for using CADtools during early conceptual architectural design process commenc-ing after the year 2000. Scholars were impressed with their excellentcapabilities especially in their advanced photorealistic visualisation ofprojects (e.g. [29] and [30]). However, doubts about the effectivenessof CAD tools in handling early conceptual design stages started almostconcurrently with their appreciation in literature. For instance,Lawson [26] ironically calls these tools “Computer Aided Drafting”rather than design. Other researches (e.g., [42]) doubt the usability ofsuch tools stating that although the CAD media have had a hugeimpact on the effectiveness of design groups, there are still char-acteristics of designing which are exclusively related with freehandsketches. A study by Kwon et al. [25] attributes this inadequacy to thelimitation of intuitive sketching capabilities of the CAD software.Therefore, they posit that conventional CAD tools might not be muchdesired during the conceptual phase.

Yet, due to increasing globalisation of design projects, designersare finding conventional sketching tools inadequate for handlingengineering parts of the design process. For instance, a researchhighlighted in [9] describes an exemplary evidence of such globalproject in the design of a Boeing 777. Its design involves 10,000designers in 238 teams scattered worldwide across 17 time zones. Wehereby note the stark differences between technical engineeringdocumentations versus the intuitive conceptual design documents.Therefore, the transitional and iterative conceptual phase is apotential knowledge-loss period [18,21] that is identified in a productdevelopment lifecycle process. This supports Fruchter's [16] earlierfindings on the potential losses of tacit knowledge within transitionsof interrupted design process. Hence, our study focuses on the con-ceptual design phase when designers would start integratingsketching with engineering considerations. Therefore, we posit thata deep understanding of design process and characteristics ofcurrently used methodologies is needed to find a successful meth-odology for handling the conceptual design phase. Many researchersin this area have conducted experiments to evaluate different mediaduring a short-term design activity in a laboratory condition involvingdesign students (e.g. [4], [31], and [41]). A number of studies com-pared conventional CAD tools against the conventional sketchingmedia and concluded that manual sketching tools are superior com-pared to conventional CAD media during the conceptual architecturaldesign process. Nevertheless, such content-oriented studies haveyet to study which aspects of the utilised media improve or hamperdesign activities' quality.

We report the results of the earlier part of our case study researchby filling up the aforementioned theoretical gap. The study by [34]adopted an ethnographic approach on a long-term and real designstudio project to obtain a better understanding on how designcollaboration, design transactions and information flow character-istics between studio masters and their students are supported byavailable technologies in a design studio project. The mentionedstudy found three types of external representation modes used bydesigners: fully manual, mixed and fully digital. The study revealedthe inflexibility of traditional geometric modelling tools withinintuitive ideations. On the other hand, it also observed the short-comings of conventional manual sketching tools in articulatingdesign ideas and translating tacit knowledge into explicit knowledgein complex design problems. We include the challenges and benefitsof each visualisation method (Table 1) during the conceptualarchitectural design phase as per summarised by the study.

Table 1Challenges and benefits of current design tools (Source: [34]).

Benefits Challenges

Current manualsketching tools

•Flexibility in ideation due to tangible interface •Lower capability for shifting from micro to macro level and vice versa•Ease of use •More tacit information flow walkthrough•Ease of learning •Lower details of visualisation•Ease of changing reforming the design alternative •Fragile models and documents for editing or reviewing•Ability for using different scales of drawing and trading ofbetween accuracy and clearness

•Failing to add and control more details into design alternative due toweak level of visualisation

•Maintaining design idea during design process providingthe ability to see all documents together and to compare

•Difficulty in transition of the format when being used in the other design stages

Current CAD tools •Easier documentation •Difficulty of obtaining ability to use•Capability for zooming and panning for easier walkthrough •Arduousness of I/O devices which interrupt creativity of the designer•Capability for temporally omitting an object or group of objects •Losing consistency of spaces due to lack of ability to control ubiquitous design idea

in an artistic way•Capability for undoing undesired changes•More detailed, realistic, and elaborated perspectives due to highcapability of visualisation

980 R. Ibrahim, F. Pour Rahimian / Automation in Construction 19 (2010) 978–987

3. Research methodology

We now extend the analysis on the aforementioned ethnographyfindings from [34] to further identify the characteristics of anoptimum external representation tool for the conceptual architecturaldesign phase among novice designers. We posit that if we couldimprove the design representationmethodology in design studios, wewill be able to expedite streamlining digitalisation procedure indesign practice of the future. The overall case study research has beenfocusing on the novice designers' collaboration culture working onconceptual design projects and the support value of the externalrepresentations during conceptual design stages. Since the case studymethod has proven to be useful for testing theoretical models by usingthem in real world situations in social science, psychology, anthro-pology and ecology fields, we are guided by Yin [46] in designing thecase study research methodology for our study. We detail the fivecomponents of the case study research methodology for thisparticular part of study in the following discussion.

3.1. Research questions

The research question for the latter part of the case study is: whatare the supportive characteristics and weaknesses of current externalrepresentation tools during the conceptual design stage?

3.2. Proposition

Knowledge plays a tacit role during the conceptual design process[18]. This tacit and ambiguous behaviour of knowledge is thestimulator acknowledged as the support for the “reflective practitio-ner's” creativity and artistic idea creation [38]. The creative amplifi-cation should not be hindered since searching for form and shape isthe principal goal of the designer during the conceptual phase ofdesign [5]. Many studies have shown that due to inadequate I/Osystems to support intuitive idea creation during the conceptualarchitectural design phase, current CAD modelling software is not yeta better alternative to replace the conventional sketching tools[4,10,25,31,41]. Hence, there are still parts of a design which arehandled by freehand sketches [42] while most other parts are beingdone digitally. This transition is known to interrupt the continuity of adesign process [25]. Taking into account the existing gap betweencurrent precise manufacturing-oriented modelling software andintuitive conceptual-based design media, this study proposes thatneither traditional manual sketching media nor conventional CADtools are good design interfaces to handle the whole design process inarchitectural education. Therefore, we posit that an alternative designmedium is needed.

3.3. Unit of analysis

Synthesis strategies of expert architects differ from those belong-ing to novice designers [3]. In this research, the unit of analysis is a2nd year architectural design studio at a local university comprising38 students and four studio mentors. We excluded data belonging toone student who had discontinued her study. Taking into account thenature of the building project that they examined, we used judgmentsampling method [24] to choose the sample population among dif-ferent studios at a design school. Homogenous conditions wereexpected since all subjects had undergone the same architecturaltraining at the design school and they were all at the same designlevel. The design project that the studio undertook was a handicraftarcade (defined as a purposed clustering of similar shopping stalls) fora town in the East Coast of Malaysia. The design objectives given to thestudents include: 1) developing design ideas through a systematicdesign methodology, 2) incorporating socio-cultural aspects in thedesign, 3) applying the knowledge acquired in the precedent study inthe design of a new building and 4) applying the knowledge of naturalventilation and daylight in designing a building. The gatekeeperduring the data collection phase was the Coordinator Studio Master ofthe selected design studio.

3.4. Criteria for linking data to proposition

This section explains the artefact and protocol analyses part of thestudy conducted for linking data to the proposition. Details are asfollow.

3.4.1. PurposeArtefact and protocol analyses were used to determine the sup-

portive characteristics for external representations during conceptualdesign stage. These techniques have proven useful to study problem-solving processes in design studies [7]. These research methods aresuggested for analysing and measuring designers' cognitive and com-municative actions instead of using subjective self-reports such asquestionnaires [23]. Scholars recommended them because of theiraccuracy in analysing the designers' actual design behaviours ratherthan the design behaviours that they would have perceived them-selves. However, since the type of media may affect a specific designtask, we resorted to evaluate the quality of these effects by classifyingand analysing the designed artefacts. Some scholars [41] supportartefact analysis because, although studying design artefacts mightfail to extract the detailed cognitive activity employed by a designer,researchers could still obtain an insight into the overall strategiestaken during the design decision making process. Scholars believethat analysing the designed solution would prevent researchers from

981R. Ibrahim, F. Pour Rahimian / Automation in Construction 19 (2010) 978–987

disrupting the design process while providing ample opportunity tocollect larger amount of data in a shorter period.

3.4.2. Types of protocolTwo types of protocols were classified based on the reporting

method: “concurrent protocols” and “retrospective protocols” [23].Concurrent protocols are collected during the task and utilised whilefocusing on the process-oriented aspect of designing. Within this“think aloud” technique, designers were requested to verbalise theirthoughts as they work on a given task [12]. Retrospective protocolsare employed focusing on the content-oriented aspects of design[11,15,37]. Here, designers are asked to describe their past thoughtsafter their tasks were completed. In normal retrospective protocolanalysis method, video recordings of the designers' actions providevisual cues on their past thoughts [23]. Hence, we also recordeddesigners' communications (i.e. retrospective protocols) in the formof communicational protocols and design artefacts since the thinkaloudmethod is not applicable in a collaborative design situation [23].

3.4.3. Coding schemeWhen utilising the protocol analysis, [4] recommend the estab-

lishment of a coding scheme for inspecting both the design activitiesand the designers' responses to any media transition. We used andadopted an existing coding scheme from [39] about media transitionwhen developing the subsequent dependent variables and indepen-dent variables. The utilised dependent variables are: 1) quality of thesolution, 2) certainty regarding the correctness of the solution, 3) totalsolution time and 4) experienced difficulty in design problem solving.These variables are dependent on the design methods (i.e. indepen-dent variables in this study) that designers adopted. This paper willexamine the first to third dependent variables while the fourthdependent variable has already been addressed in [34].

3.4.4. Data sourcesEach student was required to submit a complete set of schematic

design documents that include a site plan, all floor layouts, fourelevations, two sections, some perspectives and physical model(s) atthe end of the project. As per recommended by Kim and Maher [23],we also videotaped the designers' actions which provided visual cuesfor their past thoughts. Additionally, we recorded the designers' con-versations in the form of communication protocols and design arte-facts since Kim and Maher [23] also reminded that the think aloudmethod would not be applicable in a collaborative design situation.We also obtained data by requesting all students to declare their own

Fig. 1. A mixed method project in which the solution emulating natural forms whil

allotted time for each part of their project in a structuredquestionnaire. Finally, we relied on data extracted from thetranscribed sessions.

3.4.5. Evaluation/analysis criteriaWe assessed the quality of the solution (the first dependent

variable) by examining each design artefact in term of its degree offitting into our established design criteria where “1” is for “too poorfitness” to “5” for “so good fitness”. We asked the studio mentors forguidance on how they assessed the criteria for quality of the solutionfor grading purposes. The assessment criteria for quality of the solutioncomprise six quality components:

Q1 Development of design ideas through a systematic designmethodology. For this first quality component, each artefactwas entitled 1 to 5 points if it adopted the following strategiesin their design proposals respectively: 1 point for suggestingthe solution suddenly, 2 points for suggesting a solution usingpublished design ideas (such as those found in design maga-zines) or already existing in any contextual form, 3 points forsuggesting the solution emulating natural forms, 4 points forevolving the solution based on particular socio-cultural contextand functions of building and 5 points for evolving the solutionemulating natural forms while reflecting the particular socio-cultural context and functions of the building. Fig. 1 is anexample that receives 5 points.

Q2 Incorporating socio-cultural aspects in the design. For this secondquality component, each artefact was entitled 1 to 5 points ifit adopted characteristics of local context and architecture:proportions, forms, materials and architectural details. If theartefact has none of these characteristics, it gets 1 point.However, if it adopts all four characteristics, the artefact gets 5points. Fig. 2 is an example that receives 5 points.

Q3 Design of site plans. For this third quality component, eachartefact was entitled 1 to 5 points if it adopted the followingstrategies in their design proposals respectively: 1 point forhaving isolated elements in site plan, 2 points for having dis-tributed elements in the site plan, 3 points for touching oroverlapping elements in the site plan, 4 points for enclosing orconnecting elements in the site plan and 5 points for con-tributing elements in the site plan. Fig. 3 is an example thatreceived 5 points.

Q4 Design of façades and cross-sections. For this fourth qualitycomponent, each artefact was entitled 1 to 5 points if it adopted

e reflecting the particular socio-cultural context and functions of the building.

Fig. 2. A fully manual project considering proportions, forms, materials andarchitectural details of traditional context.

982 R. Ibrahim, F. Pour Rahimian / Automation in Construction 19 (2010) 978–987

the following strategies in their design proposals respectively:1 point for having arbitrary façade design regardless of anyholistic design idea and with no harmony among components,2 points for having façades consisting of good components orgood composition among them but no good evidence of aholistic design idea, 3 points for having façades with goodcomponents and composition among them but missing strongskylines and profiles that reflect a ubiquitous design concept, 4points for having façades exposing good evidence of holisticdesign idea with a strong composition among components andstrong skylines but with not good enough components and 5points for having façades with good components and goodcomposition among them, illustrating strong skylines andprofiles that reflect a ubiquitous design concept. Fig. 4 is anexample that receives 5 points.

Q5 Masses and perspectives. This fifth quality component considersboth interior and exterior design of the building, thereforehaving many factors contributing to the final product. Eachartefact was entitled 1 to 10 points if it adopted the followingstrategies in their design proposals respectively: 1–2 points if itis an ordinary model that fulfills the program requirement, 3–4points if the physical model and exterior perspectives illustratethe buildingmass in 3D, 5–6 points if the physical model showsthe building mass in 3D and perspectives to identify thematerials of buildings or landscape and personage, 7–8 points if

Fig. 3. A fully manual project comprising of contributing elements in site plan.

the physical model shows the buildingmass in 3D, perspectivesto identifymaterials of buildings, and landscape and personage,and 9–10 points if the physical model shows the building massin 3D, perspectives to identify the materials of buildings, land-scape and personage, and interiors including lightings, materi-als, details, etc. Fig. 5 is an example that receives 10 points.

Q6 To apply the knowledge of natural ventilation and daylight indesigning a building. For this sixth quality component, eachartefact was given 1 to 5 points when students adopted char-acteristics of green architecture: 1 point for using proper grillesand the other elements and materials on façade to controlpenetration of sunlight into the building or using thermal brakematerials (for example, wood), 2 points for controlling the ratiobetween open and close spaces for achieving better wind flowwithin spaces besides the first method, 3 points for using aterrace, veranda around the building, or canopies to controldirect sun radiations besides the two mentioned methods, 4points for using vertical void spaces (for example, atriums) forsupporting natural ventilation within the building besides thethree mentioned methods and 5 points for using appropriategreen area and landscape for controlling sun radiations besidesthe four mentioned methods. Fig. 6 is an example that receives5 points.

To determine the confidence level of novice designers or certaintyregarding the correctness of the solutions (the second dependentvariable), we relied on two major criteria. The first criterion dealtwith the level of details applied in every design artefact while thesecond one focused on the level of consistency among all designdocuments (e.g. plans, cross-sections, elevations, etc.). The latterperused how much details of one document are reflected in otherdocuments (e.g. the landscape shown in the site plan is repeated inthe perspectives).

The total solution time (the third dependent variable) wasmeasured using hours for each design activity. Technically it wasimpossible for us to videotape all 38 students' full design scenarios.However, we managed to record every student's artefact evolutionduring the semester for the final artefact analysis purpose. Yet, due tolimited individual recording sessions (there were 16 sessions in total)we obtained data for the total solution time by asking all students todeclare their own time spent for each part of the project in a struc-tured questionnaire. Finally, in evaluating the experienced difficulty indesign problem solving (fourth dependent variable) we used subjectiveprotocol evaluation for evaluating the design activities as reported in[34]. Ref. [34] evaluated the activities of each design mode byreviewing video transcriptions and perusing a student’s progressionand scheduled strategy during design problem solving to determinehis or her scores for each activity for validation.

3.5. Criteria for interpreting data

We used the Statistical Package for Social Sciences (SPSS) softwareversion 15 to quantitatively analyse the extracted data from designartefacts' analysis. We considered the quality of the solution, certaintyregarding the correctness of the solution and total solution time con-structs verified and accepted if their confidence coefficients are atleast at the 95% interval.

4. Results

Based on the reported ethnography study [34], the three dominanttypes of sketching modes—i.e. fully manual, mixed and fully digital—used by the students and their studio mentors guide us to divide ourcases into three groups respectively. Group 1 (fully manual mode—FM) used only traditional sketching tools and abstract modellingmethods, Group 2 (mixed mode—MM) started the design using

Fig. 4. A mixed method project having façades with good components and good composition among them, illustrating strong skylines and profiles that reflect ubiquitous designconcept.

983R. Ibrahim, F. Pour Rahimian / Automation in Construction 19 (2010) 978–987

traditional methods but later continued the process utilising CADmodelling tools and Group 3 (fully digital mode—FD) started thedesign in CAD environment and continued finalising the design withit. The design outputs were analysed according to the assessmentcriteria inmeeting the four design objectivesmentioned in Section 3.3.We did not consider the final grades for the design studio coursefor each student because we expected progressive improvement bystudents and other factors that may influence their total marks.

The results of the artefact and protocol analyses conducted on thedependent variables (quality of the solution, certainty regarding thecorrectness of the solution and total solution time) and the independentvariables (fully manual, mixed method and fully digital modes) arereported in the following sections. Table 2 summarises the results ofexternal representation characteristics we found in terms of theirfrequencies, mean differences and standard deviations between thethree types of sketching among the 37 students.

4.1. Quality of the solution

Due to confirmed normality assumptions for distribution of ourdata, we conducted one way ANOVA test to examine whether there isany significant difference among dependent variables belonging toall independent three groups or not. As highlighted in Table 2, thereis a significant difference in the quality of alternatives created bymembers of each group [n=37; ƒ=13.3876; ρb0.001]. The designsolutions by subjects using mixed traditional sketching and CAD

Fig. 5. A fully digital project providing perspective of the building mass in 3D, perspectivelighting, materials, details, etc.

modelling tools (MM) produce significantly higher solution qualitycompared to the other two groups (FM, FD) [(MM)–(FM): D=0.43564; ρb0.05/(MM)–(FD): D=0.91319; ρb0.001]. On the otherhand, the entirely manual sketching subjects have significantly highersolution quality compared to those subjects who solved the problemcompletely in the CAD environment [(FM)–(FD): D=0.47755;ρb0.05].

4.2. Certainty of the correctness of the solution

Significant difference has been found between the certainties ofthe correctness of the solutions of FM subjects compared to thesubjects of the other two groups [n=37; ƒ=64.353; ρb0.001]. TheMM and FD groups had used 3D prototyping techniques to ensurethat various design parts fitted and matched together, but the FMgroup (FM) was not quite successful in convincing novice designersin this regard [(FM)–(MM): D=-1.42143; ρb0.001/ (FM)–(FD):D=-1.54143; ρb0.001]. Results in Table 2 illustrate that there is nosignificant mean difference being observed when we compare thecertainty of the correctness of the solution between the FD and MMgroups respectfully [(FD)–(MM): D=0.12000; ρN0.05].

4.3. Total solution time

Results from the survey show no significant difference amongthe three groups regarding their total time taken for creating their

s to identify materials of buildings, landscape and personage, and interiors including

Fig. 6. A mixed method project comprising of all five methods for thermal comfort considerations.

984 R. Ibrahim, F. Pour Rahimian / Automation in Construction 19 (2010) 978–987

respective solutions [n=37; ƒ=2.012; ρN.05]. Referring to Table 2,the FM group had spent slightly longer time in developing their designalternatives compared to subjects who solved the problem entirelydigitally (FM) or those who had access to both traditional and CADmethodologies (MM)[(FM)–(MM): D=−170.07; ρN0.05/(FM)–(FD): D=−88.49; ρN0.05]. However, we note that the total solutiontime for novice designers using both manual and digital is slightlylower compared to the time taken by others who had solved thedesign problem completely with CAD support [(MM)–(FM): D=−81.57; ρN0.05]. We propose that the time reduction is caused by thesubjects' ability to trade-off between accuracy and the time efficiency.

4.4. Experienced difficulty in design problem solving

Since it was impossible to control the protocol size for such a hugeobservational data in term of time and number of groups from [34], weresorted to applying subjective protocol evaluation on this dependentvariable for the observed groups. The subjective protocol evaluationuses the narrative stories transcribed from the recorded videotapeswhich indicate that subjects who had utilisedmixed designmediawereable to pace their design processes with considerable less difficultycompared to subjects from the other two groups. The same subjectswere observed able to manipulate free hand sketches—as externalrepresentation tools—to solve design problems faster and easier. Theywere also able to use computational capabilities for solving theircommunicational problems either within design situations or withother designers. Results also indicate that the FMgroup had slightly lessdifficulty in producing design solutions compared to the FD group.

Table 2Comparison of the mean difference and standard deviation for quality of the solution, certa

Dependentvariable

Solution quality Certainty of th

(I) Dominanttypes ofsketching

N MeanStd. Dev

(J) Dominanttypes ofsketching

Meandifference(D=J− I)

Sig. MeanStd. Dev

(J)typske

Fully manualmode

14 3.3776 Mixed mode −.43564* 0.029 2.3786 Mix0.47416 Fully digital

mode.47755* 0.025 0.4023 Full

modMixed mode(MM)

13 3.8132 Fully manualmode

.43564* 0.029 3.8000 Fullmod

0.33698 Fully digitalmode

.91319*** 0.000 0.3674 Fullmod

Fully digitalmode(FD)

10 2.9000 Fully manualmode

−.47755* 0.025 3.9200 Fullmod

0.43670 Mixed mode −.91319*** 0.000 0.3765 MixTotal 37 3.4015

0.54590Among allgroups

f=13.3876*** 0.000 3.29460.8161

Amgro

*pb0.05 (significant difference), **pb0.01 (very significant difference), ***pb0.001 (absolu

From the artefacts and protocol analyses' results, we found that inthe case of novice designers, the best external representationmethodology among the three sketching media—FM, MM and FD—isthe one mixing both manual and digital tools. Our observationsand indicative results illustrate that neither manual sketching toolsnor CAD software are the better media for current conceptual designcommunications. We found that design semantic gets lost whenmanual design fails in articulating explicit design idea while designcreativity diminishes when using arduous CAD software. The resultssupport our earlier proposition that an alternative design medium isneeded to handle the whole design process in architectural education.Based on the listed weaknesses and strengths of each medium inTable 2, we posit that there should be another medium that supportsintuitiveness of manual sketching tools on one hand, while it alsoenhance visualisation and walkthrough capabilities of CAD toolssimultaneously. Its digital format is also expected to technically pro-vide faster documentation process when in use. We would like topropose this alternative medium to fill the gap between creativeexperimentation and precise manufacturing-oriented modelling tosupport an integrated conceptual architectural design process.

5. Validation

The case study research uses the four-step validationmethodology[46] to validate its findings. The constructs and internal validationswere obtained from studio mentors who were key informants duringthe ethnography data collection (described in [34]). For externalvalidity, this study compared the resulting artefact and protocol

inty regarding the correctness of the solution and total solution time.

e correctness of the solution Total solution time

Dominantes oftching

Meandifference(D=J− I)

Sig. Mean Std.Dev

(J) Dominanttypes ofsketching

Meandifference(D=J− I)

Sig.

ed mode −1.42143*** 0.000 466.64 Mixed mode 170.07 0.127y digitale

−1.54143*** 0.000 229.85 Fully digitalmode

88.49 0.600

manuale

1.42143*** 0.000 296.58 Fully manualmode

−170.07 0.127

y digitale

−0.12000 0.739 168.78 Fully digitalmode

−81.57 0.656

y manuale

1.54143*** 0.000 378.15 Fully manualmode

−8.49 0.600

ed Mode 0.12000 0.739 262.72 Mixed mode 81.57 0.600ong allups

f=64.535*** 0.000 382.97226.39

Among allgroups

f=2.012 0.149

tely significant difference).

985R. Ibrahim, F. Pour Rahimian / Automation in Construction 19 (2010) 978–987

analyses' findings with existing design theories that wewill present inthe next section. Finally, it used retrospective artefact and protocolanalyses [11,15,37] to analyse the procedures of novice designers'cognition and collaboration as well as the parts and changes in theperformed activities (mentioned in Section 3.4) for reliability validity.

6. Discussions and recommendations for future studies

This paper reports the second part of a case study research tounderstand the collaborative patterns for changing design media bynovice designers during the conceptual design process. We hadfurther traced and compared novice designers' collaborative protocolsfound in an architectural design studio setting. Our earlier paper [34]reports three design collaboration modes: fully manual, mixed modeand fully digital. It also found novice designers havingmajor problemswhenworkingwith current design tools whether they are analogue ordigital. The artefact and protocol analyses in the subsequent part ofthe case study research continue to support that there is no clearmedia winner between traditional sketching method and conven-tional CAD software. This latter study affirms the inflexibility of tra-ditional geometric modelling tools against intuitive ideations innovice level. The analysis of artefacts on the quality of the solution offully digital projects [(MM)–(FD): D=0.91319; ρb0.001/ (FM)–(FD):D=0.47755; ρb0.05] supports the inflexibility proposition. More-over, by analysing the certainty of the correctness of the solution, wehave equally observed the shortcomings of conventional manualsketching tools for further articulating design ideas [(FM)–(MM):D=-1.42143; ρb0.001/ (FM)–(FD): D=-1.54143; ρb0.001]. Weposit that novice designers have difficulties in turning tacit knowledgeinto explicit knowledge while working with both manual and digitaltools. The aforementioned findings lead us to consider proposing thedevelopment of an alternative design tool that would support bothintuitive idea expression besides the precise manufacturing-orientedmodelling and effortless design walkthrough.

The results of our artefact and protocol analyses show that majorbarriers with conventional sketching design tools when designingcomplex design procedures were mainly due to their shortcomings inadvanced visualisation for communication purposes as the designprocess progresses. These inadequacies were very particular fornovice designers who are not expected to have fully developed theirmanual sketching skill yet. Current conventional sketching toolscannot be replaced directly with current geometrical CAD modellingtools since our case study found that certain intuitive characteristics ofconceptual design processes cannot be supported by existing CADsoftware. Among these characteristics include the need for integrationof cognition and action—epistemic actions [13]—which are almostimpossible with arduous I/O devices of current CAD tools. On theother hand, the study saw the restricted visualisation capabilities ofcurrent manual sketching tools.

Three narratives scenarios from the ethnography observationprovide support to our arguments. In the first scenario, a femalestudent and her studio master were discussing about the student'sfloor plan lay-out. It was an example synonym to the metaphor of“walkthrough”. Even though they were only looking at a piece ofpaper, their gestures and behaviours were evidence of their immer-sion in a solution space. The studentmoved her hand on the paper andmentioned, “This is my first floor plan, and building is double story[showing some parts] and these parts will be covered with secondfloor.” Pointing to a part of the solution space, the mentor asked thestudent, “Is this your open space?” The student showed a broader partand said, “Yes, this part is open space.” Notice that neither responsecould confirm whether both parties were imagining the same space'simage. In fact, all the mentioned happenings virtually took place intheir imagination thus providing evidence of insufficiency of mediumwhen attempting to articulate their individual tacit knowledge.Hence, we identified such situation in design collaboration as a

technological source for knowledge miscommunication betweenmultiple parties.

In another technological medium scenario, the same team wasconsidering an appropriate structural system for the designed solu-tion space. The mentor showed the order of columns placed by stu-dent and said, “These are untidy. You should rearrange them in abetter order.” Then he tried to evaluate the load carrying system of thebuilding. Yet, the information presented on the paper was insufficient.At this point, both subjects had put the conventional drawings asideand used other media to improve information transactions. To simu-late the relationship of two perpendicular spatial elements, the men-tor vertically put his pencil on the table and horizontally attached hisfingers to it. Then he asked the student, “Is this the design that youmeant?” The student shook her head slightly and said, “No”. She triedto explain the system that was in her mind afterwards. First, she drewa small section on the corner of her drawing paper and asked, “Is itclear?” It was unfortunately not sufficiently clear. Suddenly, the stu-dent referred to her 3D physical model to help her clarify her designidea. She used the physical model to explain themechanism of its loadcarrying system. When viewing the physical model, both changedtheir physical positions (from standing to sitting) and also the orien-tation of the model (by turning it) frequently in order to see themodel from varied viewpoints. These actions recall the idea ofzooming and rotating facilities in much digital modelling software.The removal of some parts from the physical model in order to see theinner parts (as per managing the layers in CAD) eventually led to thephysical model's destruction.

In the third technological medium scenario, a very interestingstory happened by a student working with a popular modelling soft-ware, SketchUp™. We found SketchUp™ has some advantages overthe other CAD modelling software with its user-friendliness but wealso noticed that even SketchUp™ is not sufficient to be used inconceptual architectural design works. Our observations identifiedthat this deficiency is due to the lack of the intuitive I/O devices. Weobserved the studentworkingwith this software formore than half anhour. In the major part of the observation, the student was strugglingto overcome the shortcomings of software instead of doing architec-tural decision makings as the software uses intangible and 2D inter-face for manipulating 3D objects. Yet, there were some advantages aswell. For instance, she was able to look at her virtual model from everydistance or every viewpoint that she desired. Moreover, undoingundesired changes was much easier in comparison to those whichhappened within conventional systems. It offered much designencouragement since she can easily reverse the undesired solution.

In summary, the three narratives provide evidence that due toinherent characteristics of conventional design tools, designers stillhave difficulties in communicating their design intentions in com-plicated design tasks. Nevetheless, the study finds CAD tools havingpotentials to facilitate better communications in complex problemsolving situations. In taking into account the increasing tendencyfor substituting 3D manual modelling techniques with digital tech-nique, we would like to additionally put forth recommendations byseveral scholars who are calling for the development of an alterna-tive intermediate 3D sketching tool that could bridge the precisemanufacturing-oriented engineering design tools with the intuitiveexperimental-based conceptual design media.

A study by [27] proposes the use of some design methodologieswhich designers can swiftly produce a 3D prototype to exemplifythe 3D object they have in mind. Others (e.g. [25]) have consideredsimilar factor in order to improve computing performance for expe-diting the progress of the conceptual phase into the remaining designstages—i.e. applying the digital format as early as possible after theuse of such analogue conventional tools. Our study supports theseearlier findings as we had documented the best outcomes elicitedfrom mixed method strategy—initiating a quick conceptual sketchfollowed by detailed conceptual analysis.

986 R. Ibrahim, F. Pour Rahimian / Automation in Construction 19 (2010) 978–987

The proposed alternative intermediate 3D sketching tool must beable to overcome the obvious shortcomings of conventional sketchingtools—most particularly the intangibility of I/O devices in conven-tional CAD, the lower capability of visualisation and unmanageabledocumentation in traditional sketching media. Taking into accountthe weaknesses and capabilities of both current manual and digitaldesign methodologies, we now discuss if we could develop a 3Dsketching prototype with available VR-based 3D modelling applica-tions. A study by [25] illustrates the use of digital sketching systemswhich could develop 3D models on the computer by drawing directlyin a 3D space in a natural and quicker manner. A second possibility iscreating surfaces by moving a hand—while wearing a special glove(data glove)—through space in a semi-immersive 3D display andinteraction environment [36]. The third option is wearing a head-mounted display (HMD) with a head tracking system which effort-lessly supports a 360-degree full immersion in the design environ-ment. The fourth alternative is using haptic technologies whichfacilitate force feedback and vibration senses to fortify the tangibleuser interfaces (TUIs). Such sense of touch quality is missing in digitalsketching we had observed in the design studio. In fact, we haddocumented an obvious gap between the designer and the designartefact when they were not able to grasp and touch their designedvirtual models.

We conclude the two-part case study research with results recom-mending filling the obvious gap by using TUIs. As opposed to a simpletime-multiplexed technology used in ordinary input devices (e.g. amouse), the key advantage of TUIs is the space multiplexing inputtechnology which enables control of various functions at differenttimes [23]. In fact, Fitzmaurice [13] also relates the effects of suchinterfaces to the quality of motor activities, where he defines episte-mic or pragmatic actions for classifying designers' motor activities.He describes epistemic actions as those taken to reveal hidden infor-mation or difficult for mankind to compute mentally. He believes thephysical activities would help people perform easier, faster and morereliable on internal cognitive computation such as when one usesfingers when counting. Fitzmaurice further posits that the epistemicactions can improve cognition by; firstly, decreasing the involve-ment of memory in mental computation (termed space complexity);secondly, decreasing the number of mental computation steps(termed time complexity); and finally, decreasing the rate of mentalcomputation error (termed unreliability).

In summarising the aforementioned recommendations, thispaper establishes empirical evidence to support the potentialutilisation of VR techniques for design communication during con-ceptual design at the novice level. The proposed development of 3Dsketching in VR will act as the transitional process between analoguedesign and digital procedures hence allowing novice designersgreater flexibility between precise manufacturing-oriented model-ling and intuitive experimental-based ideation. While initial workswould focus on teaching architectural design in design schools,future studies may look into how transdisciplinary teamwork [20]could use the 3D sketching methodology in professional practice. Indoing so, such proposed 3D sketching in VR could bridge novicedesigners' understanding about interchanging epistemology be-tween tacit and explicit design knowledge concurrently from indi-viduals to inter-organisation team members. This interchange isfurther supported by migrating from vector-based conventional CADtool towards object-representation authoring tools for matching upwith Building Information Modelling (BIM)—the generation andmanagement process of building data during project life cycle [28].Integrating the conceptual design to BIM standard is expected tohelp designers demonstrate the entire building life cycle includingthe processes of construction and facility operation. This furthersupports the tendency towards digitization of design process forfuture analysis and process integration. Consequently, amplifyingdesign semantics among inter-organisational team members

throughout a project development lifecycle supports [22] proposalon the need to further enable and support information flows amongdynamic project's team memberships. In view of the aforemen-tioned findings, we would like to recommend further study ontechnology medium as a measure for establishing explicitness andreach in explaining the knowledge contingency factor that [22] isproposed in the design of discontinuous organisational structure.The proposed implementation of VR-based design technique in localdesign schools will support the proposal of [20] for the develop-ment of a new generation of architects in developing countries whowill be able to partner successfully with their counterparts fromdeveloped countries.

Acknowledgments

This study is part of a doctoral study by the second author at theUniversiti Putra Malaysia (UPM) which is partly sponsored byUPM's Graduate Research Fellowship (GRF). The authors would liketo acknowledge contributions by mentors and second year studentsin the ARC 3003 Architectural Design Studio Semester Two 2007/2008 at the Faculty of Design and Architecture, Universiti PutraMalaysia. We also thank the contributions of Prof. Dr. Mohd Salih B.Hj Jaafar, Associate Prof. Dr. Rahmita Wirza Binti O. K. Rahmat, Dr.Muhamad Taufik B Abdullah and Dr. Mohd Fakri Zaky Jaafar in thisstudy.

References

[1] Ö. Akin, How do architects design? Artificial Intelligence and Pattern Recognitionin Computer-aided Design, North Holland, New York, 1978.

[2] Ö. Akin, B. Dave, S. Pithavadian, Heuristic generation of layouts (HeGeL): based ona paradigm for problem structuring, Environ. Plann. B. Plann Des. 19 (1992)33–59.

[3] Z. Bilda, J.S. Gero, T. Purcell, To sketch or not to sketch? That is the question, Des.Stud. 27 (2006) 587–613.

[4] Z. Bilda, H. Demirkan, An insight on designers' sketching activities in traditionalversus digital media, Des. Stud. 24 (2003) 27–50.

[5] B. Craft, P. Cairns, in IFIP International Federation for Information Processing, in: T.Clemmensen, P. Campos, R. Omgreen, A. Petjersen, W. Wong (Eds.), HumanWorkInteraction Design: Designing for Human Work, Springer, Boston, 2006.

[6] N. Cross, Natural intelligence in design, Des. Stud. 20 (1999) 25–39.[7] N. Cross, H. Christiaans, K. Dorst, Analysing Design Activity, Wiley & Sons, New

York, NY, 1996.[8] N.L. Cuomo, A Study of Human Performance in Computer Aided Architectural

Design New York, University of New York, Buffalo, 1988.[9] A. Dietrich, A. Stephans, I. Wald, Exploring a Boeing 777: Ray Tracing Large-Scale

CAD Data, Real-Time Interaction with Complex Models. IEEE, 2007.[10] D. Dörner, Problemlösen als Informations verarbeitung, Stuttgart, Kohlhammer,

1976.[11] K. Dorst, J. Dijkhuis, Comparing paradigms for describing design activity, Des.

Stud. 16 (1995) 261–274.[12] K.A. Ericsson, H.A. Simon, Protocol Analysis: Verbal Reports as Data, MA, MIT

Press, Cambridge, 1993.[13] G.W. Fitzmaurice, Graspable User Interfaces, University of Toronto, Graduate

Department of Computer Science Toronto, 1996.[14] J. Fish, S. Scrivener, Amplifying the mind's eye, Sketching and Visual Cognition 4

(1990) 117–226.[15] N. Foreman, R. Gillett, Handbook of Spatial Research Paradigms and Methodol-

ogies, Psychology Press, Hove, UK, 1997.[16] R. Fruchter, Internet-based web mediated collaborative design and learning

environment, in artificial intelligence in structural engineering, Lecture Notes inArtificial Intelligence, Berlin, Heidelberg, Springer-Verlag, 1998.

[17] G. Goldschmidt, On visual design thinking: the vis kids of architecture, Des. Stud.15 (1994) 158–174.

[18] T.L. Griffith, J.E. Sawyer, M.A. Neale, Virtualness and knowledge in teams:managing the love triangle of organisations, individuals and informationtechnology, MIS Q. 27 (2003) 265–287.

[19] R. Ibrahim, Mitigating environmental characteristics with integrated design andautomated construction approaches for AQH development. ALAM CIPTA, Int. J.Sustain. Trop. Des. Res. Pract. 2 (2007) 11–18.

[20] R. Ibrahim, R. Fruchter, R. Sharif, Framework for Innovative Architectural-Construction Integration Design Studio for Cross-Borders TransdisciplinaryEducation. Archnet-IJAR:Intl. J. Archit. Res. 1 (2007) 88–100.

[21] R. Ibrahim, B.C. Paulson Jr, Discontinuity in organisations: identifying businessenvironments affecting efficiency of knowledge flows in PLM, Intl. J. Prod.Lifecycle Manage. 3 (2008) 21–36.

987R. Ibrahim, F. Pour Rahimian / Automation in Construction 19 (2010) 978–987

[22] R. Ibrahim, M. Nissen, Discontinuity in organisations: developing a knowledge-based organisational performance model for discontinuous membership, Int. J.Knowl. Manage. 3 (2007) 10–28.

[23] M.J. Kim, M.L. Maher, The impact of tangible user interfaces on spatial cognitionduring collaborative design, Des. Stud. 29 (2008) 222–253.

[24] R. Kumar, Research Methodology, a Step-by-Step Guide For Beginners, SAGEPublications, London, 2005.

[25] J. Kwon, H. Choi, J. Lee, Y. Chai, Free-Hand Stroke Based NURBS Surface forSketching and Deforming 3D Contents, PCM 2005, Part I, LNCS 3767, 2005.

[26] B. Lawson, How Designers Think — the Design Process Demystified, ArchitecturalPress, Oxford, 1997.

[27] F. Levet, X. Granier, C. Schlick, 3D sketching with profile curves, LNCS 4073 (2006)114–125.

[28] G. Lee, R. Sacks, C.M. Eastman, Specifying parametric building object behavior(BOB) for a building information modeling system, Autom. Constr. 15 (2006)758–776.

[29] L. Madrazo, Types and instances: a paradigm for teaching design with computers,Des. Stud. 20 (1999) 177–193.

[30] J. Marx, A proposal for alternative methods for teaching digital design, Autom.Constr. 9 (2000) 19–35.

[31] K. Meniru, H. Rivard, C. Be´dard, Specifications for computer-aided conceptualbuilding design, Des. Stud. 23 (2003) 51–71.

[32] A. Moum, A framework for exploring the ICT impact on the architectural designprocess, ITCON 11 (2006) 409–425.

[33] F. Pour Rahimian, R. Ibrahim, M.N. Baharuddin, Using IT/ICT as a new mediumtoward implementation of interactive architectural communication cultures,International Symposium on Information technology, Kuala Lumpur ConventionCentre, Malaysia, IEEE, 2008.

[34] F. Pour Rahimian, R. Ibrahim, F.Z. Jaafar, Feasibility study on developing 3Dsketching in virtual reality (VR) environment. ALAM CIPTA, Int. J. Susta. Trop. Des.Res. Pract. 3 (2008) 60–78.

[35] A.T. Purcell, J.S. Gero, Drawings and the design process: a review of protocolstudies in design and other disciplines and related research in cognitivepsychology, Des. Stud. 19 (1998) 389–430.

[36] S. Schkolne, M. Pruett, P. Schroder, Creating Organic 3D Shapes with the Hand andTangible Tools. SIGCHI 2001. WA, USA, ACM Press Seattle, 2001.

[37] D.A. Schön, The Design Studio, an Exploration of its Traditions and Potentials, RIBAPublications, London, 1983.

[38] D.A. Schön, G. Wiggins, Kinds of seeing and their functions in designing, Des. Stud.13 (1992) 135–156.

[39] M. Sch tze, P. Sachse, A. Römer, Support value of sketching in the design process,Res. Eng. Des. 14 (2003) 89–97.

[40] W.K. Seng, S. Palaniappan, N.A. Yahaya, A framework for collaborative graphicalbased design environments, The 16th International Workshop on Database andExpert Systems Applications, 2005.

[41] C. Stones, T. Cassidy, Comparing synthesis strategies of novice graphic designersusing digital and traditional design tools, Des. Stud. 28 (2007) 59–72.

[42] M. Suwa, T. Purcell, J. Gero, Macroscopic analysis of design processes based on ascheme for coding designers' cognitive actions, Des. Stud. 19 (1998) 455–483.

[43] M. Suwa, B. Tversky, What do architects and students perceive in their designsketches? A protocol analysis, Des. Stud. 18 (1997) 385–403.

[44] M.B. Waldron, K.J. Waldron, A time sequence study of a complex mechanicalsystem design, Des. Stud. (1988) 95–106.

[45] J. Wojtowicz, Virtual Design Studio, Hong Kong University Pres, Hong Kong, 1994.[46] R.K. Yin, Case Study Research: Design and Methods, New Delhi, Sage Publications,

London, 2003.