Design relevance in an industrial design engineering Wiki

Proceedings of TMCE 2010 Symposium, April 12-16, 2010, Ancona, Italy, ed. by I. Horvath, F. Mandorli and Z. Rusakc© Organizing Committee of TMCE 2010 Symposium, ISBN 978-90-5155-060-3

DESIGN RELEVANCE IN AN INDUSTRIAL DESIGN ENGINEERING WIKI

Regine W. VroomFaculty of Industrial Design Engineering (IDE)

Delft University of Technology (DUT)The Netherlands

[email protected]

Alexander M. OliemanFaculty of Industrial Design Engineering (IDE)

Delft University of Technology (DUT)The Netherlands

[email protected]

ABSTRACT

WikID, an Industrial Design Engineering (IDE) wiki,is an online initiative targeted at designers to facili-tate the finding of relevant information on the WorldWide Web. Because the users of this website havethe freedom to author and edit articles, it is neces-sary to reach consensus on which information is rel-evant to industrial design engineers. Therefore wehave investigated ‘design relevance’ in literature andhave conducted field research containing interviewswith experts. As a secondary objective we wish toprovide article-writing guidelines for users that helpthem to decide on design relevance and simultane-ously lower the existing technical barrier towardswriting and editing articles. Outcomes from the lit-erature study are descriptions of the scope of IDEin terms of products, used disciplines and design as-pects. These results were used to create a mission forWikID. We conclude that design relevance is not aproperty of information, but the situation where userexpectations are met. The data that was gatheredfrom the interviews showed commonalities betweenarticles in the same category but not between everyarticle. Therefore each category benefits from its ownarticle-writing guidelines. These were applied on thewebsite as ‘forms’ in order to validate the results.

KEYWORDS

Knowledge management in conceptualization, de-sign knowledge, design relevance, Wiki users, scopeof design engineering, semantic mediaWiki forms

1. INTRODUCTIONIn the field of Industrial Design Engineering (IDE) amultitude of sources is looked at during design activ-

ity before the necessary information is found [22]. Inrecent years the World Wide Web seems to becomean alternative to more traditional sources. A reasonnot to use the WWW is that however most informa-tion can eventually be found, searching for it can bevery time consuming.

An initiative to facilitate this process was the cre-ation of WikID, an Industrial Design Engineeringwiki that aims to offer information in a compactmanner, especially targeted at industrial designers.Whilst Wikipedia already offers a lot of this infor-mation, there is a distinct difference between WikIDand Wikipedia; Wikipedia aims to be an encyclopae-dia, and therefore tries to be as complete as possible.WikID aims to be a design tool. It offers informationin a compact manner tailored to its user group, beingthe Industrial Designers.

WikID also differs significantly from KnowledgeBased Systems [9] for designers and engineers, suchas WebCADET [14] and KIEF [29], because they useautomated reasoning to provide the user with relevantinformation. In order to enable automated reasoning,the knowledge base that is used by the system needsto be machine-readable, while the knowledge baseused by WikID only needs to be human-readable.

1.1. Problem area

In WikID, as in any community project, the activ-ity of the community plays a great role in the suc-cess of the entire project. The online tool is alreadyin a working state and contains an increasing num-ber of articles to attract potential users. Joris van ‘tEnde has already conducted a study [19] to deter-mine whether there is a potential for an active userbase and if so, how the process of creating an ac-

1069

tive community can be accelerated. We further in-vestigate his finding that collaborative websites witha clearmissionare more successful than those with-out. Recognizing that the domain of IDE can be de-scribed in various terms, it is necessary to decide ona framework to discuss it. For WikID, distinguishingIDE from other design disciplines is expected to bean ongoing debate and a mission will serve to set anddiscuss common goals, as well as serve as a referencefor a consensus on thedesign relevanceof informa-tion.

One of the essential elements for a starting wiki thatwas mentioned in the report [16] isseedposting, fill-ing the wiki with an amount of articles to attract po-tential users. However, seedposting is very time con-suming and that is why we have implemented an ar-ticle copying tool. With this tool users can copy anarticle from another wiki, most probably Wikipedia,to WikID under the GNU Free Documentation Li-cense [7]. This speeds up seedposting, but a directcopy from Wikipedia will not be in line with WikID’smission. The article copying tool therefore allows theuser to select sections of the article. We will investi-gate the possibility of article-writing guidelines thatwill help users in selecting sections of articles to im-port as well as help with the creation of new articles.

1.2. Purpose

To provide a clear mission we need to clarify twoconcepts: thescopeand depth of information inWikID. The scope, asking: ‘on what topics will weprovide information?’ is now defined as ‘topics in thefield of Industrial Design Engineering’. The depth,asking: ‘what information is applicable during thedesign process or, in other words, ‘when is infor-mation design relevant?’ has not yet been defined.We will attempt to define these concepts in an op-erational manner, to enable staff and community todetermine if a written article is in line with the mis-sion.

To provide article-writing guidelines that will facil-itate seedposting we need to be clear about how auser can write articles that are in line with the mission(namely, whether a subject falls within the field of In-dustrial Design Engineering and how to decide whatinformation about the subject is design relevant). Au-thors might intuitively provide a satisfactory answerto the first question but, in order to prevent disagree-ments within the WikID community, it deserves fur-ther attention. Someone with experience in dealingwith information in design practice will probably be

able to answer the second question, but an inexperi-enced student or someone specialized in an IDE re-lated field, but without experience in design practice,will likely be unable to answer it adequately.

WikID’s article-writing guidelines should assistusers in answering these questions. As a goal for thisstudy we have set out to provide operational defini-tions for ‘scope’ and ‘depth’ to be used for WikID’smission (goal A). Our second goal is to create article-writing guidelines to help users decide on design rel-evance (goal B). In the following section we elabo-rate on how we will achieve these goals.

1.3. Methods

To provide operational definitions for the concepts‘scope’ and ‘depth’, we have formulated the follow-ing research questions:

A.1 On what topics will WikID provide information?

A.2 When is the information design relevant?

In answering these questions, we will first perform adesk study to determine the degree to which existingtheories might already answer them. After the deskstudy we will decide whether additional research isnecessary to complement existing theory, or whetherwe have found sufficient answers to these questions.Finally, we will attempt to summarize our results in asimple model (the mission page), as an aid for WikIDusers.

In order to compile article-writing guidelines to helpusers decide on design relevance, we have formu-lated the following question:

B.1 Which article-writing guidelines assist users indeciding on the design relevance of information?

The desk study performed to answer A.2 will alsoprovide the theoretical basis to answer this question.By means of an explorative study, we will gather datafrom subjects experienced in deciding on design rel-evance in order to expand our theory with a set ofguidelines. The detailed description of the followedmethods is given in section 3. After compiling theseguidelines, we will test them with WikID’s users inorder to validate our theory.

2. DESK STUDY

This section deals with the process and results of thedesk study performed to clarify WikID’sscopeanddepth. The goal is to construct an operational defi-nition – designed for, but not limited to WikID – toenable moderators and users to determine whether ar-

1070 Regine W. Vroom, Alexander M. Olieman

Material things

bridgesculpturemountain farmlandair planethand

mineral oil

potato

houseplant

aquariumsand

milk

coffee beans

building

lathetree turret press aircraft truck

Consumer

products

Products

plastic resin

TVvacuum cleaner

gasoline

ink cartridge

aspirin

ballpointpen

inkjet printer washingmachine

detergent

garbage

bin liner

napkin

dung

garbage bin

candybeer

table

car telephone

IDE core

insurancepolicy

training

Durables

Artefacts

Material things

bridgesculpturemountain farmlandair planethand

mineral oil

potato

houseplant

aquariumsand

milk

coffee beans

building

lathetree turret press aircraft truck

Consumer

products

Products

plastic resin

TVvacuum cleaner

gasoline

ink cartridge

aspirin

ballpointpen

inkjet printer washingmachine

detergent

garbage

bin liner

napkin

dung

garbage bin

candybeer

table

car telephone

IDE core

insurancepolicy

training

Durables

Artefacts

Figure 1 Scope of the prpducts (Van der Vegte).

ticles are in line with WikID’s overall purpose. Weexpect to provide a satisfactory answer to our first re-search question and to provide a partial answer to oursecond research question in the desk study.

2.1. Scope: on what topics will WikIDprovide information?

Since WikID is an initiative of the Faculty of Indus-trial Design Engineering at the Delft University ofTechnology, the scope is connected with the currentviews of this faculty. The desk research thereforestarts by looking at their mission statement.

“The field of industrial design engineering consistsof all activities concerning methodical and creativeinnovation and the development of (parts of) con-sumer goods. The objectives of the Faculty of IDEare described in the mission statement (2004):

‘Creating successful products people love to use’

[. . . ] The Faculty of Industrial Design Engineering’sconcern is to study, innovate and improve the devel-opment of durable products and their related servicesfor people, on the basis of the balanced interests ofusers, industry, society and environment.” [17]

This tells us that these design engineers developconsumer goods, consisting of durable products andtheir related services. In this activity they balancethe interests of users, industry, society and environ-ment. To achieve their goals, they use knowledgefrom the disciplines of design, technology of mate-

rials and production, engineering mechanics, math-ematics, formgiving, technical drawing, presentationtechnique, ergonomics, product development, busi-ness & organization, marketing and research [17].

Scope of products

IDE bachelor students are taught that products whichare tangible, man-made, series or mass-produced,made predominantly for consumers and durable are‘typical IDE products’ [20], illustrated by a Venn-diagram. We refer to products that conform to allcriteria (Table 1) as the ‘IDE core’ and we refer to theborder regions of the IDE core, simply as border re-gions, see Figure 1. A more complex representationof the product scope might also be more detailed, butkeeping our goals in mind we aim for simplicity.

Table 1 IDE core criteria.

Core criteria1. the product is tangible2. it is human-made, artificial3. it is series or mass produced4. primarily for personal or family use5. durable, usage over time

Products that lie outside the IDE core, which are con-sidered to be in the border regions sometimes are sobecause it is arguable that they are intended for usein daily life or are frequently used, involving signifi-cant interaction between user and product [20]. This

DESIGN RELEVANCE IN AN INDUSTRIAL DESIGN ENGINEERING WIKI 1071

Figure 2 Matrix of disciplines and the role of engineering design (Horvath).

is the case with the turret press in Figure 1, which isnot a consumer product and therefore not includedin the core. However, consumer products that arenot durable are also often considered to be in theborder region if they are integral to the functioningof a product that is included in the core. A similarbridge might be found for certain non-material con-sumer products.

Scope of disciplines

Recognizing that the scope of products still includessome of the work of related design practices, suchas fashion design and graphic design, we choose tolook at a second scope. The problem in classifyingthis correctly is primarily due to the fact that otherdesign practices also work towards series and massproduced durable consumer products. The most dis-cernable difference between these practices and IDEis that the knowledge used to design originates fromvery different disciplines.

In a third year course in research methodology stu-dents learn about the differences between science anddesign. The slides for the first lecture state that sci-ence aims to find truth where design aims to find ef-

ficiency [2]. To explain the relationship between sci-ence and design, a chart of a matrix is given, wherevarious disciplines are organized on two polar axes,from analytic to synthetic and from abstract to con-crete. We reproduce this chart as Figure 2 Matrix ofdisciplines.

Knowledge in IDE is constructed using knowledgefrom various sciences as well as more synthetic dis-ciplines. To realize efficient product design, a transi-tion must be made from mono-disciplinary to multi-disciplinary knowledge. Therefore the field of indus-trial design engineering has an integrative approachto knowledge management. To illustrate this rela-tion, it is sketched on top of the matrix of disciplines,reproduced in Figure 2 [10].

Current structure in WikID: design aspects

In 2005 a method to structure information was stud-ied for C-DET, a precursor to WikID [23]. Vroom etal. produced a list of design aspects that is currentlyin use in modified form. The top level of this cat-egorization is noteworthy, because it is another wayof describing the domain of industrial design engi-neering. It tells us which aspects are commonly ad-


dressed during the product design process. The en-tire designaspects list can be browsed through onWikID [25]. Design Aspects is one of the three maincategories in WikID. The other two are Design The-ory and Product Domains.

Table 2 Top-level design aspects.

1. materials2. production3. ergonomics4. aesthetics5. legislation6. packing7. working principles8. energy techniques9. logistics

10. quality11. costs12. environment13. product safety14. components15. remaining aspects

2.2. Depth: when is information designrelevant?

To start a literature search for design relevance, weexpand our search terms to be context specific. Whenis “domain” “knowledge” of “industrial design engi-neering” “relevant information”?

While we are reaffirmed in the idea that it is im-portant to support designers with relevant domainknowledge [21], none of the sources found specifi-cally define design relevance. An overview of the ac-tivities of the Design Methodology Group does, how-ever, provide us with a list of authors who might havewritten about this subject [6]. Our search is contin-ued in the direction of publications by these authors,following references from and to interesting papers.

Studies by Henri Christiaans

First we discuss an overview of research conductedby Christiaans et al. at the DUT between 1993 and2000 [3]. In [5] the results of an empirical researchare presented which shows that subjects using theirinformation retrieval system used it mostly in theanalysis phase and in defining the design specifica-tions. Most subjects started with user-related topicsbut in later phases mainly searched for legal normsand standards.

A later study by Snoek, Christiaans & Hekkert [15],

studying the influence of contextual information onthe design outcome, demonstrates that outcome ispositively affected by the relative amount of con-textual information that is processed. However, thisstudy showed that the use of contextual informationdid have a negative impact on feasibility and con-structability of the designed objects. As an overallconclusion of the overview, Christiaans & Restrepo[4] state that consideration of valuable contextual in-formation is often absent in design activity, but thatit is not clear what contextual information actually isand how it can be of most use to designers.

According to [4], information access is primarily im-portant during problem structuring and secondarilyduring problem solving. During problem solving,the designer mainly uses technical data, while dur-ing problem structuring the designer needs more con-textual data, causing more complexity in informa-tion access. Because of the vagueness in writtensources dealing with contextual data, designers willoften make extensive use of human sources. On de-sign relevance Christiaans & Restrepo state the fol-lowing: ”access to information will be improved ifthe information provided is deemed by the user asrelevant, for relevance is not a property of the infor-mation itself, but an attribute endowed by the user ina certain situation”. The key to creating design rel-evant articles lies in meeting the designers’ expecta-tions of the content of the document. They conclude,that”we need a better understanding of the structureof the problem and solution spaces, as well as of theprocess of structuring problems, as it will allow fur-ther insight into the process followed by designers”.

Studies by C eline Mougenot

The first study by Mougenot et al. that we encounter,deals with differences in the use of pictures betweenprinted media and internet sources in the early stagesof the design process [13]. It tells us that“today’scomputational tools don’t allow looking for inspira-tional materials at a chosen level of abstraction”,implying that designers would like to do so. Assum-ing this is the case, it would be beneficial to stimulatediversity of inspirational materials through WikID’sarticle writing guidelines. More importantly they un-veil that in searching for pictures on the web theirsubjects share a common interest. When referringto pictures as belonging to a creative sector, 69,1%of the selected images belonged toproduct design,architectureand interior design in similar propor-tions for each designer. This could imply that in the


broader area of formgiving, it is possible to distin-guish certain areas of preference in dealing with in-spirational materials.

The second paper provides us with a clear approachtowards dealing with contextual information, pre-sented as problematic by Christiaans et al. Theypresent the creation of software named TRENDS,which allows designers to search for multimedia ina flexible way [1]. TRENDS is based on ConjoinedTrends Analysis (CTA) which is used in turn to facili-tate designers to enrich the solution space by lookingat various domains of influence while identifying for-mal trend attributes, such as shape, color and texture.While the solution they present is very interesting, itis focused solely on images and is therefore of lim-ited relevance to WikID. The methods needed to useCTA digitally also go beyond the limitations of Me-diaWiki, the software WikID is running on.

2.3. Findings

In the scope of products (Figure 1) we can discernan IDE core that is composed of products that con-form to the core criteria (Table 1). Products that donot conform to all of these criteria might also be in-teresting to industrial designers. It is expected thatauthors who wish to write about these products canargue why a certain product is considered to be inter-esting.

The scope of disciplines (Figure 2) shows that IDEmostly uses knowledge from the concrete disciplines,together with a slight preference for synthetic dis-ciplines as opposed to analytic. The design aspects[25] provide a tailored structure, crafted to the pur-pose of WikID. Yet they are unavoidably always re-ferring to the past. As the future of IDE plays out, theopen MediaWiki software provides users and staffwith the possibility to let WikID grow together withthe world it provides knowledge on.

Now, asking again on which topics WikID will pro-vide information, we can answer that topics will con-tain knowledge from certain disciplines that can beused in the design of certain products (specified by2.1). The current structure in WikID is appropri-ate and expandable, built to last. On a practicalnote, staff should be alert for top-level category ad-ditions and changes because they have important im-plications for WikID’s knowledge structure. Askingwhen information on an appropriate subject is designrelevant cannot be answered by studying literature.Thinking as design relevance as a property of the in-formation leaves the user, the person dealing with in-

formation in WikID, out of account. It is perhaps notpossible to reason our way to design relevant contentfor a large user group in one step, but using a step-by-step process, we can certainly work towards it.

To arrive at a tailored solution for WikID, we have tokeep taking into account that from the use of a wikisystem follows the need for thinking in terms of masscollaboration. No desk study will accurately predictwhat users want to read about an appropriate subject,unless it gathers direct data from WikID users. Userswill consider information to be design relevant if ithelps with problem structuring and problem solvingin their current design situation. Following the con-clusions from Christiaans & Restrepo [4], we predictthat they will primarily need contextual data and sec-ondarily need technical data to structure and solveproblems.

Providing design relevant information on appropriatetopics lies in meeting needs and expectations fromother WikID users. The designated places for thismultitude of meetings already exist in MediaWiki;they are the talk pages that are available for everyarticle in WikID. By using semantic properties inWikID in a clever way, it might even be feasible topresent an overview of requests in for example a cer-tain category using real-time data from WikID itself.A successful application of this concept might workby showing where there is a shortage and where thereis an excess of certain metadata descriptors, tags orkeywords.

3. FIELD STUDY

Following the findings from our desk study (2.3.) wecan say that in order to fill WikID with design rele-vant information, or in other words meeting currentuser needs and expectations, we need to take into ac-count the mass collaborative nature of any wiki sys-tem and allow users to communicate their needs tothe whole community. In a field study we will gatherdata from WikID users and potential users.

3.1. Subjects

Our subjects will be chosen for their experience inthe field of IDE. Subjects are active as design stu-dents, university staff or work in design practice.Most subjects are, or have been, in some way con-nected to DUT. To participate in the study subjectsneed at least a basic knowledge of the concept ofa wiki, therefore persons who are unfamiliar withWikipedia are excluded.


3.2. Equipment

Subjects perform a list of tasks on WikID andWikipedia using a provided laptop or another avail-able pc in spaces where auditory distraction is judgedby the researcher to be below a workable threshold.The researcher takes notes, records audio and pro-vides technical assistance. Both the web browser his-tory in the laptop and activity on WikID are recorded.

3.3. Procedure

Introduction

While the researcher sets up the laptop, the subject ispresented with the text from section 1. Introduction.At this point the researcher may answer any questionthe subject has about WikID. The instructions thatthe subjects are presented with are available online[4] and are discussed in this section.

User account

If the subject already has a user account on WikID,he or she logs in. If not, a user account and personalpage are created to record basic data, such as theirname and number of years of experience with IDE.

Article selection

The subject is presented with a list of wanted pages,which contains over a thousand entries of articles thatare wanted in WikID but do not yet exist. Whenthe subject has intuitively selected an article from thelist, Wikipedia (or another source) is searched for anarticle that can serve as the basis for the new articlein WikID. This step can be repeated until the subjectis satisfied with the selection.

Import wizard

The import wizard in WikID is used to copy the se-lected Wikipedia (or other) article to its new destina-tion in WikID. Then, still in the wizard, the subjectselects sections from the Wikipedia (or other) articlethat they judge to contain design relevant informa-tion. The researcher takes care to refrain from in-fluencing the subject’s decision making, only askinghim or her to verbalize considerations and perhaps topause for a moment to clarify a taken action.

Editing

Now that the sections containing design relevant in-formation have been copied to their destination arti-cle, the article can still be edited manually. Becausewe are interested in the subject’s considerations and

not their skill in editing MediaWiki source code, theresearcher may perform the edits, but only acting onclear commands that are issued by the subject. Forthe sake of duration no information is added, onlyremoved. On inquiry, the subject will be informedthat additions can be made in the next step. This stepends when the subject is satisfied with the final articleresult.

Talk page

If the subject wants to add information to the article,they have an opportunity to do so. However, they aretold that the talk page, linked by the ‘discussion’ but-ton in MediaWiki, also provides the chance to leavesuggestions for improvement. They are encouragedto note at least one mention of a section that couldbe improved and a more general mention of informa-tion that would be useful but is absent in the currentarticle.

Interview

The researcher informs the subject that we are look-ing for rules of thumb to assist users in deciding ondesign relevance. The subject is asked to look intohis or her memory for decisions made concerningdesign relevance, perhaps in the performed tasks butalso in earlier experiences, and is invited to talk aboutthe possibility of using personal rules of thumb. Ifthey succeed in stating any personal rules of thumb,they are asked if these rules of thumb might be mod-ified to be of value to a broader public.

4. RESULTSA total of ten subjects with 3 – 30 years of experiencewith IDE completed the procedure in the period fromApril to June 2009.

4.1. Gathered dataDue to the compact nature of this publication andthe volume of our data, the detailed results of theinterviews are available online [27]. The topics ofthe articles that have been added in this study are:mp3 player, walker, welding, plasma cutting, bicycle,telematics education, headphones, composite mate-rials, compact disc and manufacturing. These werechosen by the subjects.

4.2. Interpreting dataThe gathered data show that most similarity occursbetween topics in the same category. For instance,the articles named mp3 player, walker, bicycle, head-phones and compact disc are all articles in the main


category Product (Domains). And between the ar-ticles within such category, we found most similar-ity. Therefore the decisions, ideas and argumentsof our subjects were clustered into the categories ofproducts, production techniques and materials. Thecategories production techniques and materials areboth categories within the main category of DesignAspects. The results from the telematics educationand manufacturing articles were considered not to beconcrete enough for this approach. The order of theresults is derived from the order of the imported ar-ticles. Interesting comments that are not directly re-lated to our research questions are discussed furtherin the discussion section 6.

The main category: Products (domains)

The results for this main category Product (domains)are based on the similarities in the articles about mp3player, walker, bicycle, headphones, and compactdisc.

Considered relevant

The following information parts in articles aboutproduct (domains) are considered to be design rel-evant:• History (recent predecessors, development)• Operation (technology, structure)• Subdivision (different variations on the main

product)• Standards (per region)• Usage (risks, personal, professional, environ-

ment)• Visual description (components)• Competitive products (benefits and limitations)• How the product functions (physics, construction)• Regional name (British equivalent)• Recent patent

Considered irrelevant

The following information parts in articles aboutproduct (domains) are considered not to be designrelevant:• Controversy (media debate)• History (first commercial mp3 player)• Expired patent• Popular culture (rock group)• Social and historical aspects

Requested by subject

The following information parts would be useful ac-cording to the subjects, but are missing in the currentarticle:

• Quantitative information on usage• Specific information on operation per subtype• Taxonomy• Link to trustworthy website for trends and social

aspects

Category Production techniques

The results for this category Production Techniques(within the main category Design Aspects) are basedupon the similarities in the articles about welding andplasma cutting.

Considered relevant

The following information parts in articles about pro-duction techniques are considered to be design rele-vant:• State-of-the-art• Subdivision• Possibilities and limitations• Safety issues• Costs of the technique• Physics formula


The following information parts in articles about pro-duction techniques are considered to be design rele-vant:• History• Details of the devices used with the technique• Costs of the devices

Materials (composite materials)

In the current study only one article about materialshas been added (about composite materials). How-ever in a former study [18] the category materials(which is one of the categories in the main categoryDesign Aspects) was the main subject. Therefore wewill use the information found in the current studyto check the results of the specific form for materialsthat is already present within WikID.

Considered relevant

The following information parts about compositematerials are considered to be design relevant:• Introduction (from history)• Possibilities for molding and tooling• Mechanical properties• Examples of subtypes• Application in products


Figure 3 The scope of products in WikID’s mission.


The following information part about composite ma-terials is considered not to be design relevant:• Reference of historical literature

Requested by subject• Examples of products using the material• More internal and external links

4.3. Findings

The subjects who chose to import an article with aproduct as topic broadly consider the same informa-tion to be relevant. Only one subject contradicts otherusers in stating that“WikID doesn’t seem to be theplace for background information”[13]. The sub-jects who chose a production technique as a topicalso complement rather than contradict in their ideaof design relevance. However, from the differencein relevancy between products and production tech-niques we conclude that there is no single set ofrules-of-thumb that will assist users in deciding ondesign relevance throughout the whole of WikID.There are multiple sets possible: one for productsand one for each of the design aspects. Followingthe field results [27], we expect that no such set isneeded for design theory, because each theory thatis applicable by designers is in principle design rele-vant.

It is possible to attach a template or form to a cat-egory in WikID. This on the one hand assists userswith deciding on article content and structure, but onthe other hand also benefits the uniformity of infor-mation in WikID. This uniformity creates possibili-ties for semantic queries, the automatic generation of

overview pages and the ability to re-mix informationin a creative manner while always keeping contentup-to-date. Section 8 deals with the creation of tem-plates and forms for products and production tech-niques. Since our results on materials (4.2.) nor addto nor contradict with our current templates and formfor the materials category [18], no revision of thistemplate is necessary.

5. VALIDATING RESULTS WITHINWIKID

Reviewing our goals (1.3.), we set out to providean operational definition for the scope and depth ofWikID and to provide guidelines to assist users indeciding on the relevancy of information. To val-idate the results of this research, a mission articleand several forms that invite users to contribute infor-mation which is considered to be design relevant bytheir peers were created on the website. The discus-sion pages for the mission and the forms will serveto evaluate and improve on their first versions.

5.1. Mission

The mission page [28] serves mainly to enable thecommunity to set and discuss common goals. It con-tains the definition of the scope, as in paragraph 5.1.Figure 3 shows a screen caption of this part of themission. Since our definition of the depth (2.3.) isstill dependent on both the user and the topic of anygiven article, it is not yet fully operational. Thereforethe mission refers to the separate forms for products,materials and production techniques based on section4.2. as partial definitions for the depth.


Figure 4 Part of ‘Form:Products’ and the resulting article ‘Waste container’ .

5.2. Forms and templates

The MediaWiki extensions Semantic MediaWiki andSemantic Forms were used to create forms that askthe user for the information that the subjects fromour field study considered to be design relevant. Thisinformation is passed on to several templates to ren-der the resulting article. Since several of the subjectshave suggested that WikID should focus more on theuse of images, an option to include an image throughthe form has been added to all forms. To illustratehow a form and a resulting article appear to users,Figure 4 is added as an example.

Figure 5 shows the entire products form. The top in-put field asks the user to upload a main image and toadd a description for it. The following ten text fieldsask the user for information that was considered rel-evant by the subjects from the field study. Some ofthe fields contain even more specific instructions forauthors. For instance, in the history field they are

asked: “please focus on recent history and develop-ment”. At the bottom a category is selected in whichthe resulting article is placed and the possibility isgiven to mark the article as infringing copyright, be-ing incomplete or containing false information. If the‘watch this page’ checkbox is checked, the user willbe informed of subsequent changes to the article.

The production techniques form (Figure 6) looksquite similar, but asks users to contribute differentinformation. It also has an extra ‘used on’ field atthe top, which is used to generate semantic data ofwhich materials and products can be used with thisparticular technique. This enables for instance usinga semantic query to generate a list of all productiontechniques that can be used with wood.

The existing form for materials is included as an ap-pendix. All templates that are used by the forms canbe found online in Category:Building blocks [24].


Figure 5 Form:Products.

6. DISCUSSIONIn the course of this research several interesting is-sues that were not directly related to our researchquestions have presented themselves. They are dealtwith in this section, because they either evaluate anexisting aspect of WikID or present a topic for futurestudy.

Figure 6 Form:Production techniques.

6.1. Navigation

During the field study the subjects often displayedsub-optimal navigational skills and required assis-tance from the researcher. They had most troublewith navigating from an article to parts of the cat-egory structure and from an article to other articlesin the same category. In the implementation of theaforementioned forms (5.2.) the option arose to addnew navigation templates that will automatically be


Figure 7 The result of ‘Template:Related articles’ on the ‘Soldering’ article.

added when the form is used to create a new article.The CategoryTree extension was used to place a cat-egory tree under the main image, as seen in Figure 4.At the bottom of the article a template is inserted todisplay neighboring articles and their main image, asseen in Figure 7.

6.2. Import wizard

The articles that are created when a user imports anarticle from Wikipedia with the Import wizard in-clude links to articles and images that do not ex-ist on WikID. Often the links point to articles thatshould not be in WikID at all, yet they are addedto the ‘wanted pages’ list. This is confusing whenusers check the wanted pages to see where their con-tribution is needed. Therefore an extension [8] is cur-rently being written by Alex Olieman and Roel Ob-dam to edit the links on these pages, offering func-tionality to unlink the words and upload missing im-ages.

6.3. Transclusion

As one subject has indicated [27], importing articlesfrom an external source is only a good option if it isexpected that the article will soon be improved on bythe community. It is better to transclude the articlefrom the source until original user content is writtento replace it. Transclusion includes (part of) a doc-ument into another document by reference. In thisway information from external sources can be em-bedded into internal articles, ensuring that the con-tent remains as up-to-date as its original source. Thisis especially valuable in the starting phases of an on-line community because only a small number of ar-ticles can be improved by user contribution, leavingmost imported articles as a duplicate of their exter-nal source until the community grows. To enable thisfunctionality in WikID the wiki configuration needs

to be changed [12] and an extension is needed to pro-vide extra functionality [11].

7. OVERALL CONCLUSIONS

The detailed findings of our desk study are found inparagraph 2.5. and those of our field study are foundin sections 4.3. and 5. For the sake of overview werepeat our research questions in this section, each be-ing followed by a compact conclusion.

A.1 On what topics will WikID provide information?

Industrial designers use information from existingdisciplines, showing a preference for disciplines thatare both concrete and synthetic, such as the appliedsciences and engineering. WikID should provide in-formation on these topics as long as care is taken tomake them accessible to industrial designers as op-posed to for instance physicists.

Within IDE, knowledge from existing disciplines isalso used to construct new, IDE specific knowledge.This knowledge is by default relevant to designersand WikID should always provide information on it.

WikID can also provide information on products thateither serves as inspirational material for design ac-tivity or that is the result of design activity. To dis-cern typical IDE-products from other products, a listof criteria was made (Table 1).

A.2 When is the information design relevant?

‘Design relevancy’ cannot be seen as a property ofinformation, because only the user of the informationcan deem it to be relevant. On a per-case basis, the

information is relevant when user expectations arefulfilled.

Our field study shows that, although different design-ers have different expectations or preferences whendealing with information, if topics reach a certain


level of similarity the expectations between design-ers are also more similar.

When applying this into guidelines to help users de-cide which information is design relevant, it is im-portant to acknowledge the differences between mosttopics and therefore create guidelines for a specificgroup of topics (being a category in WikID) in whichexpectations are similar.

B.1 Which article-writing guidelines assist users indeciding on the design relevance of information?

Several WikID users have already commented thatlack of time is the greatest barrier towards contribut-ing. Instead of listing guidelines (prescribing), wehave chosen to use forms, inviting users to contributecertain pieces of information.

The guidelines that are based upon our field studyhave been put into application on WikID. Users areasked to evaluate them and improve on their first ver-sions.

Based on the results of this study, there is nowa framework for discourse on design relevance inWikID. This enables authors to write about topicsthat identifiably belong to the domain of IndustrialDesign Engineering. The creation of forms willcause there to be more similarity in the structureof articles within a category, so the information inWikID will better meet the expectations of the usersand will be more accessible to the user group.

REFERENCES

[1] Bouchard, C., Omhover, J.-f., Mougenot,C., and Aoussat, A., (2008), “TRENDS: AContent-Based Information Retrieval Systemfor Designers”, Design Computing and Cogni-tion (pp. 593 - 611). Springer.

[2] Christiaans, H., (2008), “College 1, febru-ari 2008”, Blackboard.tudelft.nl, courseOnderzoeksleer-Q3-ID3701.

[3] Christiaans, H., and Restrepo, J., (2000), “In-formation processing in design”, DRN, (pp.63-73).

[4] Christiaans, H., and Restrepo, J., (2004),“Problem structuring and information access indesign”, Journal of Design Research.

[5] Christiaans, H., and Van Andel, J., (1998),“Information Processing and Storage Duringthe Design Process”, In E. Frankenberger, P.Badke-Schaub, & H. Birkhofer, Designers-

The Key to Successful Product Development(pp. 233 - 248). London: Springer-Verlag.

[6] Dorst, K., (1995), “The Design MethodologyGroup”, Retrieved february 24, 2009, from de-signresearch.nl:http://www.designresearch.nl/PDF/DRN1995Dorst.pdf

[7] Free Software Foundation, (2008), “GNU FreeDocumentation License”, Retrieved December9, 2009, from The GNU Operating System:http://www.gnu.org/copyleft/fdl.html

[8] futureplanetwiki.org, (n.d.), “Future PlanetWiki:Extensions”, (n.d.), Retrieved July 31,2009, from futureplanetwiki.org:http://futureplanetwiki.org/wiki/Future PlanetWiki:Extensions

[9] Hayes-Roth, F., and Jacobstein, N., (1994),“The State of Knowledge-Based Systems”,Communications of the ACM, Vol. 37, No. 3(pp. 26-39).

[10] Horvath, I., (2005), “On the development ofengineering design science”, In I Nobuyaki, YTakeda & T Kanada (Eds.), Proceedings of the1st International Conference on Design Engi-neering and Science - ICDES 2005 (pp. 5-10).s.l.: JSDE.

[11] mediawiki.org, (n.d.), “Extension:LabeledSection Transclusion”, Retrieved July 05,2009, from mediawiki.org:http://www.mediawiki.org/wiki/Extension:LabeledSectionTransclusion

[12] mediawiki.org, (n.d.), “Man-ual:$wgEnableScaryTranscluding”, RetrievedJuly 5, 2009, from mediawiki.org:http://www.mediawiki.org/wiki/Manual:$wgEnableScaryTranscluding

[13] Mougenot, C., Bouchard, C., and Aoussat,A., (2007), “A Study of Designers’ CognitiveActivity in Design Informational Phase”, Pro-ceedings of the 16th International Conferenceon Engineering Design.

[14] Rodgers, P.A., Huxor, A.P., and Caldwell,N.H.M., (1999), “Design Support Using Dis-tributed Web-Based AI Tools”, Research in En-gineering Design, Vol. 11, No. 1 (pp. 31-44).

[15] Snoek, H., Christiaans, H., and Hekkert,P., (1999), “The Effect of Information Typeon Problem Representation”, 4th InternationalDesign Thinking Research Symposium on De-sign Representation. Massachusetts Institute ofTechnology.


[16] TU Delft / IDE / Department O&S., (2008),“IDE Graduation Manual”, Retrieved February13, 2009, from Algemene informatie over afs-tuderen: http://www.tudelft.nl/live/pagina.jsp?id=bb068469-cc1a-4e86-9cbc-c35188a48237&lang=nl&binary=/doc/IDE%20Graduation%20Manual%20december%202008.doc

[17] TU Delft / IDE / Department M&C., (2008),“Programme of Industrial Design Engineer-ing”, Retrieved February 14, 2009, fromhttp://www.io.tudelft.nl/live/pagina.jsp?id=35596315-9921-4a55-92b3-cdbb0798cde7&lang=en

[18] Van ’t Ende, J.J., Jelierse, R., and Olieman, A.,(2008), “Towards an Industrial Design Engi-neering Wiki”, Delft University of Technology:internal report.

[19] Van ’t Ende, J.J., Vroom, R.W., and Kooijman,A., (2008), “Creating a community base forWikID”, Delft University of Technology: in-ternal report.

[20] Van der Vegte, W.F., (2008), “Products in Ac-tion - IDE1030”, presentation slides of the firstlecture. Delft, Delft University of Technology.

[21] Vergeest, J.S., Opiyo, E.Z., Kuczogi, G.,Wiegers, T., and Horvath, I., (1999), “Require-ments for highly interactive system interfaces”,CIRP, (p. 4). Delft.

[22] Vroom, R., Kuiper, C., and Wassink, M.,(2004), “Knowledge Search Problems AndStrategies Used By Design Engineers”, 4th In-ternational AED 2004 Conference.

[23] Vroom, R., Pape, R., and Van Vulpen,E., (2006), “Design Engineering informationsearch through design aspects tree”,AED.

[24] WikID.eu, (n.d.), “Category:Building blocks”,Retrieved from WikID.eu: http://wikid.eu/index.php/Category:Buildingblocks

[25] WikID., (n.d.), “Category:Design aspects”,Retrieved March 9, 2009, from WikID:http://www.wikid.eu/index.php/Category:Designaspects

[26] WikID.eu, (n.d.), “subject0409”, Re-trieved April 22, 2009, from WikID.eu:http://www.wikid.eu/index.php?title=User:Amolieman/subject0409&oldid=4722

[27] WikID.eu, (n.d.),“User:Amolieman/results0409”, Retrievedfrom WikID.eu: http://www.wikid.eu/index.php/User:Amolieman/results0409

[28] WikID.eu, (n.d.), “WikID:Mission”,Retrieved from WikID.eu:http://wikid.eu/index.php/WikID:Mission

[29] Yoshioka, M. et al., (2004), “Physical Con-cept Ontology for the Knowledge Intensive En-gineering Framework”, Advanced EngineeringInformatics 18 (pp. 95-113).


APPENDIX: FORM:MATERIALS


1084 This page is deliberately left empty.

Design relevance in an industrial design engineering Wiki

Documents

Transcript of Design relevance in an industrial design engineering Wiki