Material Objects in Design Process

!"#$%&'#$&($)*+#,-*'$./0#1+2$-3$+"#$4#2-53$6,&1#227$8$9&:;*,-2&3$&($!
<&$4#2-53$9='+=,#2$*3>
?&<$!"#@$9&3+#3>$<-+"$8=+&:*+-&3
8=+"&,A2B7$C*+",@3$?#3>#,2&3
D&=,1#7$D1-#31#E$!#1"3&'&5@E$F$?=:*3$G*'=#2E$G&'H$IJE$K&H$IE$AD;,-35E$LMMNBE$;;H$LJMOLPQ
6=/'-2"#>$/@7$D*5#$6=/'-1*+-&32E$R31H
D+*/'#$S%T7$http://www.jstor.org/stable/690180
811#22#>7$LNUVWUIVVN$LX7VV
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=sage.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
http://www.jstor.org
The Roleof MaterialObjectsin the
DesignProcess:A Comparisonof Two
DesignCulturesandHowTheyContend
withAutomation
Kathryn Henderson
TexasA&MUniversity
This article comparestwo culturesof engineeringdesign, one flexible and interactive,

the other rigid and hierarchical.It examinesthe practices of design engineers who use
a mixtureof paper documentsand computergraphics systemsand contraststhese with
the practices of workers reengineeringtheir own work process and its technological
supportsystem,usingpredesignedsoftware.Based on the ideafromactornetworktheory
that objectsparticipate in the shaping of new technologiesand the networksthat build
them,the studyreveals that (1) design culturesare intrinsicallytied to the way in which
theirrepresentationsare constructedbecause suchrepresentations-sketches,drawings,
prototypes-are the heart of design work; (2) such design tools can engage or restrict
participationin the designprocess; (3) politics in theform of managementprerogatives
can be builtinto a design tool, influencingthe range of creativityallowed and innovation
accomplishedin a given sociotechnologicalsetting.
This articlecomparestwo culturesof engineeringdesign, one flexible and

interactive, the other rigid and hierarchical.It examines the practices of
design engineers who use a mixture of paper documents and computer
graphicssystems and contraststhese with the practicesof workersreengi-
neeringtheir own work process and its technologicalsupportsystem, using
predesignedsoftware. Design culturesor styles are intrinsicallytied to the
way in which each constructs representationsof their ideas. Such repre-
sentations-sketches, drawings, prototypes-are the heart of design work
and constitutethe space in which ideas are defined, refined,and negotiated.
Comparisonof the mannerin which differentdesign tools both engage and
restrictparticipationin diversesettingsrevealsthe importancenot only of the
tools, butalso of the culturethey helpbuild.The studyalso shows thatpolitics
in the form of managementprerogativescan be built into a design tool,
Science, Technology,& HumanValues,Vol. 23 No. 2, Spring 1998 139-174

? 1998 Sage PublicationsInc.
139
140 Science, Technology,& HumanValues
influencingthe rangeof creativityandinnovationin a given sociotechnologi-

cal setting.
Actor networktheory (Callon 1986a, 1986b; Latour 1987; Law 1987a,
1987b) lets us treat design documents-sketches, drawings,diagrams-as
participantsin the interactionsthat constructnew technologies. Such docu-
ments are actantsenlisted into and constructedby a heterogeneousnetwork
of people and things which jointly contributeto the constructionof a new
technology. Visual representationsare also the holding ground for collec-
tively negotiateddesign ideas.Theyserve as recruitingdevices andgatekeep-
ers. As I have shownelsewhere,visualrepresentations,includingprototypes,
used in the design of new technologiesareconscriptiondevices, objects that
both enlist and constrain participationin design. Anyone who wants to
participatein thedesignprocessmustdo so by interactingwith them.Because
they representa design group's negotiatedideas, and hence its distributed
cognition, they also structurehow work gets done in groups (Henderson
1991, 1995a). Consequently,their form and format affect the nature of
interactionsand the culture of the designing group. Art historian Alpers
(1983) has shown that how members of a given culture see the world is
reciprocallyinfluencedby the way they renderthis world. In his discussion
of visualizationand cognition,Latour(1986, 9-10) refersto this relationship
as "definingboth what it is to see and what there is to see." I have shown,
similarly,that engineeringdraftingconventionsand theircontext help con-
structthe visual cultureof engineers,thatin turninfluencesthe conventions
of renderingthat world (Henderson1995b). Clearly,the implementationof
new representationaltools (such as computergraphicssystems) changesthis
culture,while the existing conventionsinfluence how the new tool may be
used (Henderson 1993, 1994, 1995b). Because visual representationscan
facilitateor restrictgroupinteraction,they are primaryplayersin the social
constructionof the design culture or design style of the designing group,
whetherits membersareperiodartists,mechanicalor otherdesign engineers,
reengineeringworkers,or any othertypes of designers.
This study adopts the STS perspective that views technology and its
design as both socially constructedand society shaping.This perspectiveis
compatiblewith design studies thatcomparedifferencesand similaritiesin
variousdesign cultures.Whereasthe role of visual renderingsin design has
been often acknowledged,science and technology studieshave been paying
attentionto the role of visual representationsnot only in design but also in
laboratorywork1and in the heterogeneousengineeringof technology (Fer-
guson 1976, 1992; Hindle 1981;Mukerji1984; Latour1986,1992; Fyfe and
Law 1988; Bowker 1988; Henderson 1991, 1995a, 1995b;Turnbull1993;
Gormanand Carlson 1993). Historical (Vincenti 1990) and ethnographic
Henderson/ Role of MaterialObjects 141
studiesby engineersthemselves(Bucciarelli1994) also documentthe impor-

tance of visual documentsin technology design development.Observations
of engineeringpractice (Ullman, Stauffer,and Dietterich 1987; Bly 1988;
Tang and Leifer 1988; Henderson1991, 1995a) have pointed out that more
is going on in sketchinganddrawingthanthe workingout of a single design.
Fromthe psychological perspective,Arnheim,who has long recognizedthe
importanceof visual thinking(1969), has more recently (1993) pointed out
the dialectic natureof sketches for the individualworkingout design ideas.
From the sociological perspectiveadoptedhere, such a dialectic is manifest
at the grouplevel when engineersanddesignersuse interactivesketches and
drawings on white boards and paper to work out and negotiate various
perspectives and to draw in, literally and figuratively, a wealth of tacit
knowledge (Henderson1991, 1995a, 1998).
The new perspectivesin design studies are somewhatanalogousto those
in science and technology studies. Sociological perspectiveson science no
longer acceptthe myth of linearscientific progressnor embracethe myth of
scientific objectivityand detachmentin researchmethods.Similarly,design
studiesaremoving away frompositivistand functionalistviews thatprescribe
the one rightway to do design (such as formfollows functionideology). They
tendto advocateexaminationof the varietyof design processesandacknowl-
edge no preferredright way to do design and no single correctmeaning of
the designed object in a pluralisticsociety (Margolin 1989; Margolin and
Buchanan1995; Buchananand Margolin 1995). Accordingto Margolin,in
the continuing debate over the definition of design, more positivist views
such as those of HerbertSimon, who defined design as the science of the
artificial, compete with ideas such as those of the civil engineer Petroski
(1985), who describeddesign as the disassemblingand reassemblingof the
partsof nature,and with the perspectivesof those who suggest thatdesign is
a new formof practicalartor communication(Margolin1989). Most relevant
to this study arethe ideas of ItaliandesignerVitta,who sees a design culture
as embracing"thetotalityof disciplines,phenomena,knowledge, analytical
instrumentsandphilosophiesthatthe design of useful objectsmust takeinto
account,inasmuchas those objectsareproduced,distributed,andused in the
context of economic and social models thatare even more complicatedand
elusive" (Vitta 1989, 31). Vitta's inclusion of such heterogeneouselements
moves his positioncloser to the perspectiveof actornetworktheory,although
he emphasizesthe consumptionratherthan productionside of the network.
The overall theme that design is an activity defined by the social milieu in
which it occursis compatiblewith the perspectivetakenhereandagreeswith
those in science studies who argue that neither design nor science and
technology are value free.
The contingenciesof automationin work worlds thathad been tradition-

ally dependenton paper-basedpracticeshave also drawncross-disciplinary
attention.Researchershave shownthatin manyworksettings,practicesusing
paperareso integralto the way workis done andknowledgeis generatedthat
they cannotsimply be replacedby automatedsystems (Harper,Hughes, and
Shapiro 1991; Luff, Heath,and Greatbatch1992). Despite nationalisticU.S.
hope and hype, implementationof computergraphicsin all types of engi-
neering design-from civil to mechanical to electronic-has not led to a
smoothertransitionfrom design to manufacturing(Downey 1992a, 1992b,
1995). It has also not led to a paperlessworld,butratherto a world of mixed
practicesthatcombinethe assets of boththe paperandthe electronicworlds.
Initial ideas are often capturedon paper, whereas analysis and problem
solving are often accomplished individually and/or collaboratively using
computer-generatedhardcopies (Henderson,1993, 1994, 1998).2Hence the
flexibility of collaborative practices around visual documents is crucial,
whether the representationsare hand-drawnsketches on paper and white
boards,orcomputer-generated hardcopies,redlinedby hand.Moreimportant
thanthe differencebetweenelectronicrepresentationalformatsis the differ-
ence betweenflexible andnegotiablevisual representations(andthe situated
interactionsaroundthem) and the rigid andreified formats.Both the work-
place context and the particularrepresentationalmode used arerelevant.
The article begins with an overview of my findings from ethnographic
researchconductedin industrialengineeringdesign settingsandfocusing on
the crucialrole of visual representationsin thecollaborativelysharedknowl-
edge of team design work.I thendescribethe participationactionresearchat
a national call center and present the situated practice of one of the jobs
to be integratedthroughreengineering.Finally,I comparemy observations
of the reengineers'practicesusing printoutsfrom their company-provided
software with my earlier findings regardingdesign engineers' interactive
practices on, around, and through visual documents. The contextual ele-
ments, includingthe way automationis handled,are analyzedalongside the
choices of the mode of visual representation.
Visual Documents and Situated

Practice in Design Engineering
The visual culture of design engineering that has been built around
drafting conventions constructs the manner in which engineers see their
world. My earlier research revealed the crucial role visual documents
(sketches, drawings, schematics) play not only in the constructionof new
technologies but also in the constructionof shared and individual work

practices and knowledge (Henderson1991, 1995a, 1995b). For instance, a
case studyat a site thatbuildsturbineengines revealedthatvisual documents
wereused individuallyandcollectively as thinkingandnote-takingtools. The
importanceof sketches as individualthinkingtools is illustratedby a drafter
who hadmoved throughthe ranksto become a design engineer(still possible
in some old-line, heavy industries).When she was promotedto engineerthey
took away her draftingboard,but she askedfor it back, stating"Ican't think
withoutmy draftingboard."She describedthe way drawinghelps her think:
Whenyou do draftingby hand,a lot of the thoughtprocesscomes when you're

drawingit out. So it's like you can't think.... Youcan't sit thereandwait until
you've got it up here,becauseit comes throughdrawingit. And as soon as you
startdrawingit, you haveideasandchanges.You'reerasingit andimprovingit.
You'rejust gettinga feeling for it. Youknow,you'rejust tryingto get a size,
tryingto understandit-to get anunderstandingof the system,almostbuilding
up your self-confidenceso you know the task you've been assigned.
Then, as soon as you have something,you can take it to someone else and
say, "Look,this is what I have, how can I improveon it?"or "Did you have a
problemin this area too?" Withoutthe drawing,you know, it's just talk. But
when you startlayingit out, that'swhere,you know,80%of the problemscome
out, when you start drawing it and you start realizing what you've
got. . . you've always got somethingin hand. (Henderson1991, 159-60)
Thus, as individualthinkingtools, sketches help to captureconcepts in a

tangibleand communicableform which then facilitatesindividualand com-
munal critical analysis. Pairs and groups use sketches as an interactive
communicationtool for furtherrefiningof ideas.Collaborationover sketches
is one of the best methods for accessing and sharingtacit knowledge. On
several occasions at the same site I observed seasoned design engineers
engaged in collaborativesketching conversations.One engineer described
such activity:
I don't think you ever get two designers who just sit down andjust talk. It's
"give me a pencil, and I'll explain."Everybodydrawssketchesto each other;
[it's] "Youknow, this is whatI'm tryingto do here,let me show you." ... They
say thatthe best designersstartdrawingrightfromthe beginning.That'swhere
they makeall theirmistakes.... They'redrawingit out there.They're looking
at it. They're visually checking it and improvingupon it. (Henderson 1991,
460-61 )
Duringa study at a high-precisionmedical optics firm I observedeven more

synchronouscollaboration-two engineersdesigning a new surgicalinstru-
mentwere so engagedin sharedvisual thoughtthatthey werejointly drawing
a sketchwith one implement,each alternativelytakingthe pen fromthe hand

of the otheras they drew and talkedsimultaneously.
Moreover,throughoutmy observationsI found, as have others (Ullman,
Stauffer,and Dietterich 1987; Bucciarelli 1994), thatdesignersdo not actu-
ally follow design school protocolsthatdictateworkingfrom overall design
to finer and finer detail in a linearfashion,developing one areaat a time. In
practice,designersworkbackandforthfromgenerallayoutto specific detail.
Nor do they concentrateon one area alone and develop it. They work on
different portions of the design, all over the drawing, both serially and
simultaneously,developingit as a whole while they collaboratewith others-
a nonlinearpracticethattakes advantageof multipleresourcesand perspec-
tives by negotiatingdistributedcognition.
Productivemechanicalengineeringdesignersoften takeearly sketches to
the shop floor and show them to machinistsand welders. The combination
of the visual representationand the interactionregardingit serves both to
elicit informationthatimprovesthe design and to make explicit information
that might be ambiguous in the drawing. As objects serving to facilitate
sharedcognition, sketchescapturepertinentknowledgefrom many sources.
Cap, the senior designer on the turbinepackage project, pointed out the
crucialrole thatsketchesanddrawingsplay not only in collaborationbetween
design engineersbut in eliciting tacit knowledge in consultationwith those
in the productioncycle. He describedhis methods:
Whenyou're doing the mechanicaldesign work you have to be very awareof

all the people thatareinterfacingandneed to workwith yourdesign problems.
You've gotta leave room for the electricalpeople to put their stuff, you have
to know how to put a package togetherso it... can be built in the cheapest,
most inexpensive way that you can possibly do it. An' to do that you are in
contact a good portion of the time with other departmentsespecially, eh,
structures,which actuallyweld the thing together.
I usuallytakemy layoutsrightafterI've started,afterI do my initiallayouts,
beforethey're ever dimensionedor, you know,just be a framelayout,then I'll
go over and I'll sit andtalk with them and say, "Look,eh, here's what I wanna
do and here's the types of materialI wanna use, here's what I've gotta have
cut out." . . . And it's just a matterof: "is this the best possible way to do this
particularthing?"And usually it's peripheral,but a lot of times they'll say, ah,
"Well,gee, if you did this, over here, this would save one weld." (Henderson
1991,460)
Because he knows from experiencethatdrawingscan be misinterpreted

on the shop floor, this engineeris out on the productionfloor duringthe first
assemblyof his designs. This way he can provideadditionalvisual informa-
Henderson
/ Roleof Material
Objects 145
tion when needed. Referring to the plan, profile, and section drawings from
which the initial assembly workers must build the first prototype, Cap
acknowledges:
They still may not make as much sense in one area, so you circle a line on it
andmakea little handsketchoff to one side thereto clarifythe thing. [Pointing
to a vague areain a layout] For instance,like where its black here in this area,
they see a dotted line coming down and say "what'sreally going on there?"
I'm liable to just go off with the pencil andgo an'just makea little handsketch
off to one sidejust to makeit clearerto the guys doing it at the time. (Henderson
1991,461)
Latour (1986) points out that such cascades of sketches cumulatively build
up knowledge for the ultimate transformation from paper to machine at
centers of calculation that are necessarily centers of power. My observations
ethnographically document the larger complexity of this process as informa-
tion is accumulated in a continual progression back and forth from paper to
machine to paper until all the myriad design details reach closure through
negotiation of the multiple heterogeneous social and technical elements and
players. For instance, at the medical optics site prototypes played an impor-
tant role in the development of the new surgical tool. They were interspersed
between the iterations in the cascade of drawings that transformed the design
idea into a producible product. As the project engineer put it:
It [a drawing]helps in termsof, "Ohit looks pretty"or "Howdoes this [work],"

or "I think my finger could work really well on this." But until you actually
get your hands on it, it's still not even close. There's a big, big difference in
drawingit andmakingit. That'swherethe mainchangecomes over, when you
get the prototype.... I would say you cannotmake anythingunless you make
prototypes,urn,andunderstand it andworryoutthebugs.(Henderson,1995a,283)
Prototypes were especially useful for soliciting tacit knowledge, which was
then incorporated back into new iterations of the drawings. This eliciting of
tacit knowledge was seen as very useful play by the engineers involved:
I'm basicallythe personwho comes in and says, "Ah,this isn't workingright"

and then I say, "Let's change this and see if this will fit and get rid of that
problem."... Or I bring them into my office and play with the instruments.
That'swhen thingsget accomplished... when I'm grindingat thebit andthat's
just the next thing on the list-one of the criticalthings to know is thatif you
keep working on the thing all the time it doesn't get any better .... You can
work super hard on it. But it just doesn't improve. Going every time with a
fresh approach,with a fresh look and not being afraidto change it is just as
important.(Henderson1995a, 286)
Such play with the prototypewas greatlyadvancedby the engagementof

others,especiallynew membersof theprojectteamwhojoined the designers'
meetings when the new tool neared production. The research and devel-
opmentteam's strategyof handinga prototypeof theirinstrument,designed
for physicians'use, to newcomers(myself included)and soliciting feedback
about the appropriatenessof its size, shape, smoothness, and comfort of
function served not only to garnertacit knowledge (Polanyi 1958, 1967;
Collins 1974). Politically,it also workedto facilitatesupportfor the design-
ers' "technologicalframe"in Bijker's terms (1995), "jumpingon the band-
wagon" in Fujimura'sterms (1988), or "buying into the project" in the
commercialterms of my corporateinformants.Hence prototypesserved as
devices for eliciting and capturing-conscripting-tacit knowledge and
as devices for conscriptingsupportnetworks. They share the capacity of
paper documentsto be read on several levels at once, as boundaryobjects
(Star 1989; Star and Griesemer 1989).3But this capacity too has political
possibilities becausemultiplereadingscan bothfacilitatecollaborativework
and contributeto conflict amongcollaborators.As Bucciarelli(1994) points
out, the shareduse of visual representationscan help participantsnegotiate
differences,given theirdifferentexperientialbackgroundsand expectations
for the design. Negotiationcan also be facilitatedwhen participatingengi-
neersreadthe design drawingsfromthe perspectivesof theirspecializations
(for example, from the perspectiveof fluids flows, electrical, or structural
engineering). However, diverse orientationscan also potentially lead to
out-and-outconflict when readingsare based not only on differentspecialty
orientationsand backgroundexperience, but also on departmentalpolitics
and distrustbetweenresearchanddevelopmentandmanufacturing.My data
from the practiceof design engineersshow thatbecause of the potentialfor
such conflict, visual documentsplay a gatekeepingrole as agentscontrolling
who is permittedto have inputinto the new design. Designers at more than
one site told me that they did not show early iterationsto superiors,so that
certain decisions were alreadyclosed by the time managementinput was
given. The way in whichdesignershadto contendwithmanagementpreroga-
tives in the reengineeringprojectwas very differentandmuch moredifficult
to circumvent.Before discussingthe work of the designersin thatproject,it
will be necessary to understandthe context of the reengineers'contrasting
design culture.
Reengineering a Call Center
The site of this studywas one of severalcall centersthatprovidessupplies

and service for a multinationalcompanythatmakescopy machines.41was a
memberof a team of sociologists engaged in action researchto participate
along with workers in reengineeringtheir work. The project plan was to
integratethreespecialtyareas-service, supplies,andbillinginquirycall-taking-
into one moregeneralistcall-takingjob. Initially,calls for service came in on
separate800 toll-freelines to telephonerepresentativesin the threerespective
areas.Underthe reengineeringinitiative,suchcalls wouldcome in on a single
800 line, requiringthe TeleRepto become knowledgeablein the multitudeof
policies and practicesfor all three areas.New softwarewould facilitatethe
integratedfunctions.In the first phase of the researchour team documented
existing practiceso thatthe reengineeredcomputertools andwork processes
would be groundedin actualways of doing work.
This phase of our researchpaid close attentionto the total context of the
TeleRep'sinteractionswith people, work setting, and machines.Our obser-
vations focused on the interdependenceof conversationstructure,the use of
paperdocuments,the dictatesof the technology,the role of tacit knowledge,
and the physical involvementof the TeleRep'sbody. We found the driving
forces for interactionandthe locus of the knowledgeto be multiple,shifting,
and mutating.Knowledge was not singularlyconstructedinto the "smart"
system that supportedthe work alone, nor only into the heads of TeleReps,
singly or collectively. Nor was it located in the posted documentsalone, nor
only in the TeleRep'sfingertipsor visual memory;it was in all these places,
elicited interactivelyand in consort,sometimesorchestratedsimultaneously
and other times sequentially(Whalen,Henderson,and Whalen 1995). This
has been called distributedcognition by Hutchins (1991). Hutchins now
terms it cultural cognition (1995). Suchman(1987) and Lave (1988) call it
situatedpractice; it is knowledge in practice.The importantpoints are that
(1) the knowledge that TeleReps command was complex and distributed
throughouttheir tacit, technical,and physical work environment,including
collaborationwith peers; and that (2) the menu-drivengraphicmethods the
companyprovidedfor representingand redesigningthe work, coupled with
a structured,management-drivenquality assuranceprocess, workedagainst
capturingthis complexity in a mannerthatwould allow creativeinnovation.
In all three areasof specialization(service, sales, billing inquiries)to be
combinedby reengineering,the TeleRep,almost cyborglike,sits suspended
between one machine-the telephone-and another-the computersystem.
Headphones connect her or him to the computerizedtelephone system,

fingers hovering at-the-readyover the keyboard,and eyes set to scan the
computerscreen.Also withinat-a-glancerangearenumerouspaperdocuments
andpad andpencil. On the surfacethe worklooks simple:answerthe phone,
place the requestin the system. However,thejob is made more complex by
the myriadvariationsin types of service,purchase,andlease contracts, time
zones, prioritycodes, supplies used by various models of copiers,quantity
discounts, availability,and special offers, to namejust a few details that the
TeleRep must maneuver.To keep trackof much of this informationalong
with its constant updates and the coding for the various screens of the
database of an aging system, TeleReps, especially new ones, paper their
space with posted and bounddocuments.I mentionthem here only as part
of the context;the detailsof theiruse is an articlein itself. To understandfully
the complexity and knowledge involved in actually doing this work, I
participatedin the standardtrainingclasses in serviceandsuppliessales given
to new employees.I spenttwo weeks takinglive calls in the CustomerService
Call Center.Because thatjob was generallyregardedas the simplest of the
three to be integrated,I will draw on that area to illustratethe attentionto
detail andtacitknowledgeinvolved in the work of the TeleRepsat its locally
perceived minimum.
The CustomerService Call Centersegment of the convertedwarehouse
in which the national call center resided was broken up by clusters of
two-personworkspaces,separatedby cubicle walls risingabouta foot above
the work surface.Higherwalls compartmentedworkareastowardthe rearof
the large space, makingit quieterthanthe ongoing call-takingchatterof the
main floor. Sets of clocks on threewalls of the centralspace displayedtime
zones across the continentalUnited States, while red LED screens above
them reportedthe numberof callers waiting on hold and how long they had
been waiting. A light on the telephoneconsole on each desk also indicated
urgency,turningfrom green to yellow to red as the caller queue lengthened.
The console's standardtelephone keypad was amplified with additional
largernodes thatlit to indicatewhetherthe TeleRepwas on an 800-line call,
an internal,company line, or an outside line; and in "wrap-up,""ready,"or
"idle."
The telephone system is computer controlled, but a human monitor
surveys the display,showing who is off the system, or, if on, which type of
line they are using. All of this is recordedand checked against statistical
norms and expected on hours. The results are posted monthly, a form of
employee monitoringand discipline. This is anotherexample of the politics
built into artifacts (Winner 1985, 1995). Even though management has
recentlyinformedemployees thatthese statisticsare less importantthanthe
quality of service they give customers, TeleReps tend not to believe this
rhetoric, because the monitoring devices are built into the technology and
have not been removed. My own experience supports the TeleReps' percep-
tions, as shown in this account from my field notes while I was taking calls:
Mara,my trainerand the floor facilitatorcomes by and asks thatI sign off the
telephonesystem when I am takingnotes and asking questions because I am
staying in wrap-upmode on the phone too long, causing the computermoni-
toring system to highlight the designationfor my telephone line connection,
upsetting their process. Once signed on to the phone, TeleReps are really
hooked in like partof the machine,every action monitored.I ask a neighbor
TeleRepif I would have been reprimandedfor this, were I a regularemployee
and she says, "Yes."5
Although the work of taking and recording service calls can eventually
become monotonous, learning the tacit knowledge required both to use the
technology and to manage the callers takes time and experience, as illustrated
by this excerpt from my first day on the phones:
I plugmy headset into thephoneconsole in my assignedcubicle,next to Sandy,

an experiencedTeleRep.She helps me log on, but a screen message indicates
a PC problemso she rebootsthe machine.Wechoose the appropriateprograms
from the programmanager,Rhumbaand FWSS. I type in "logon impf," hit
enter, then the ctlc22 button twice, as I have been instructed,but the screen
tells me: "your logon is alreadyin use." We have to call the trainerback. It
turnsout someone else has been given this numbersince I used it yesterday.I
get a new numberand we do it all over again.Finally I am readyto take calls.
I punch the ready buttonon the phone console. A recordedvoice tells me
"incomingcall." I respondas I have been taught,"[companyname] customer
service, this is Kathryn,How may I help you?"The customeridentifieshimself
and his company,saying he wants to place a service call. I say, "MayI have
your serial number,please?"He is used to the protocol,has his serial number
handy and rattlesit off so fast I have to ask him to repeatit so I can type it. I
make a couple of typos and the numberturnsred. Finally when I have the
correctnumberit turnsgreen andthe datascreenfinally comes up. I verify the
caller's and his company'sname and addressbut Sandy has to remindme to
verify the telephone number.Its field is buried in the middle of the screen
among many others and I missed it in my visual scan. The customertells me
he has lines and streaks on his copies so I enter "poorcopy quality"in the
problem field. Sandy tells me later I could have just entered"cq."I tell the
customerI will place the call andhe hangs up. I hit the enterbuttonto register
the service order.The purchaseorderfield turnsred and I have to enterN for
"no purchaseorderrequired"before the system will let me enter the order.I
hit enteragain and the screen for solving machineproblemscomes up. Sandy
has to help me get out of it. I should have hit the returnkey, located rightnext
to the enterkey. I also forgot to verify the customerhours, importantfor the
service technicianto know, Sandy remindsme. This field is in the uppermost
rightcomer of the screen,which I also missed in my visual scan as I readdown

the page, vertically,to verify the name and address.Sandy also tells me not to
use the hyphen in telephonenumbersbecause it sometimes messes up pager
calls-a glimpse of TeleReptacit knowledge.
I take two more simple requestsfor service calls but keep forgettingto ask
the caller for the condition of the machine:totally down, partiallydown, or
routineservice check, for which I should enter a 0, 1, or 2 in the appropriate
field. A new call comes in, but when I ask for the serialnumberthe caller says
they areplacingme on hold andwill transferto theircopy room.Too muchtime
passes,butnowI cannotgetoutof hold,on thephonesystem.Itrypressingthehelp
button and get a recordingstatingthat all my calls are on hold, and to call a
supervisor.Herline is busy andso is Sandy,so I sit feeling dumbandfrustrated,
afraid to punch any more buttons that might make mattersworse. Finally,
Sandy is free and tells me to pressthe blinkingbuttonof the several lit up on
the phone console. This is my first piece of informationthat the buttons are
active not static, more tacit knowledge absorbed.
The opening protocol for any service call requireseliciting the machine
serial numberfrom the caller to access the databaserecords.The database
was designed in the 1970s. Some but not all TeleRepshave personalcom-
puterswith a Windowsinterfacethroughwhich to access the database.Many
terminals,however,have the old green-on-blackdisplay,andall information
in the database, even with new PC color choices, appears in old style,
squarish,difficult-to-readcomputerprint.Whateverthe customermay wish
to convey about a nonworkingcopier, no informationcan be recordedfor
potential action until the machine serial numberis recited and enteredinto
the system. If the entered serial number is valid, the computer system
proceedsto the next screen.The informationfields on this screenareso dense
they fill the entirepage, often leaving only one- and two-digit spaces to be
filled in. The visual field is so crowded that it is easy for a novice to miss
details. Only after the TeleRep verifies the name, address, and telephone
numberof the corporatecustomer,along with the name of the client contact
personandcompanyhoursof operation,is the customerto be askedwhatthe
TeleRep can do for them. Whatever a caller replies about their machine, only
a two-line problemfield is availablefor the description,so the TeleRepmust
usuallyrewordand shortenthe description.If a tiny field at the bottomof the
screenis markedMC for maintenancecontract,the call is enteredby simply
pressingthe appropriatekey, andthe TeleRepinformsthe customerthattheir
call has been placed. If the field is markedTM for time-and-materials,the
TeleRepmust informthe customerthatthey will be billed a certainamount
for the first half hour and anotheramountfor each subsequentquarterhour,
dependingon the size and complexity of theirparticularcopier and whether
the customerwants the service call duringor after regularbusiness hours.
The most current billing rates are posted on paper nearby so that TeleReps
can verify that they are quoting the most recent price updates. But not all calls
are so straightforward, as these examples from my field notes show:
The next call is a legitimateENOF [equipmentnot on file]. The serialnumber

the customerhas providedturnsred when I attemptto enter it. I have double
checked for typos on my partand had the customergo back to theirmachine
to check the numberthey gave me against that on the copier's silver serial
numberplate. This in itself is an ordeal.Many times customersdo not know
theirserialnumberandmust go look for it. It is not in the same place on every
machine. Each TeleRep has a small but thick reference volume with a line
drawingof each model of the company'scopier,typewriter,and fax machines
and their accessories. An arrowpoints to the location of the serial number,
which is sometimes inside the machine. Sandy can tell customerswhere it is
on the most common models without looking it up; she has worked for the
company for twelve years. I write down the company and caller's name,
address,telephone and serial numberson a pad of paper.Sandy says to wait
until we come back from lunch to handlethis because we will have to make a
phonecall andhave someoneget backto us so we can tell the customerwhether
the call is billable or covered on a service contract.
The data screen for the next caller has "XFS in conversion,customersaid
was convertingfrom time and materialsto ser.cont."in the billing field. We
never discussed anythinglike this in trainingso I ask Sandywhat I should do.
She says to put the service call throughbecause it was alreadychangedat the
top of the screen and was internalanyway.I am learningthatthe literalrules
we weretaughtin trainingas the one-and-only-way-things-are-done areindeed
more open to interpretation,even in an automatedsystem.
One of the forms of tacit knowledge I had to develop as a TeleRep was

learning to see the data screen. I had to discipline my visual scan to include
the small two-digit fields pertinent to what I was doing at the time while
glancing over other fields not relevant to the customer on the line. This meant
training my gaze to include the far upper-right corner of the screen where the
clients' hours were posted instead of reading down the screen, vertically, as
if it were a list. Learning to scan all the relevant fields took some practice:
I get call from an elementaryschool thathas "donot service, no renewalPC,"

an 800 number,and a name to call, so I tell the woman that there is some
problem on the account and she needs to call the 800 number.I have now
trainedmy visual scan to notice the billing field at the bottomof the screen as
well as those things I have to verify. Sandy tells me a lot of schools have
problemsgetting service becauseof billing problemsbecause they often have
really old equipmentwithoutservice contractsor have to get purchaseorders
from theirdistrictor state.
I get severalcalls thatI handleOK,thenone fromthe Stateof Massachusetts
Departmentof MotorVehicles on which I missed the "PC[purchasecontract]
required"until afterthey hung up. Sandy says to just put "if billable " in that
field. I get a call from CarmellaEastinat InvestmentBankersof Americawho
said she had called yesterdaybut received no ETA[estimatedtime of arrival]
andfelt she shouldhave by now, Sandytells me that[readingthe screenbetter
thanI] the service technicianhas alreadybeen out there.
A bit after2:00 p.m. [CST] Sandyremindsme to ask customerson Eastern
StandardTime whetherthey want to pay for after-hourscalls. This involves
not only noticing where and in what time zone the customeris located and
whetherthey have a service contract,but also if the service contractcovers
after-hourscalls.
Although I had been through training, I felt very dependent on my paper

cheat sheets, which listed the various computer codes to move through the
database or protocols for exceptional, but common, breaks from simple calls,
like the one below.
I do my first CCE [customercourtesyescalation],the processfor impatientor

iratecustomers.The caller was irritatedthatthe technicianhad not yet come.
I designateit a"+1"ratherthana "+2"becausethe tone of the customer'svoice
is not out-and-outanger,but mild irritation.I must send two messages from
the screen, one to the area service technician supervisor and one to the
designated technician.With my cheat sheets near-byI feel confident that I
know the process but then get confused about how to designate the service
team numberbecausethe exampleused in trainingwas unclear.Workingwith
the datascreenratherthanmy memoryandthe cheatsheet,I realizethatthough
we have to enterthe same message twice, it is the second box on the message
screen thatactuallysends the message to the technicianvia employee number
when I put"CCE+ teamandsectionnumber"into the CCP[customercourtesy
process] screen.I make the usual mistakeof using the enterbuttonto end this
which means the technician got paged twice. Sandy reminds me I should
have used the pflO key, adding,"Thenshe would not have been paged twice."
The service area was local so Sandy knew who the technician was by the
employee number.
My second day on the phones, paper cheat sheets at hand, I felt more
self-assured and resolved to try to use the on-line system (CCA; customer
call assistance) designed to help customers fix simple problems with their
copy machines themselves without having to place a service call. I had heard
Sandy give instructions to customers from memory. I would be dependent on
the system. After a delay because the whole system went down, forcing us to
scurry around getting paper documents to record calls, I finally had a chance
to try out the technology-assisted helper role:
I take a call from an Army Base in their CareerResource Mgt. Center.The

caller says their machine is "jammingin the output area"and give[s] me a
complicatedset of hours,differentfor each day of the week. In additionto the
PO [purchaseorder]numberthe caller wants to give me a very long "BPA"

number.I write all the hoursand the BPAnumberdown on my tablet and ask
thecaller,"Areyou willing to trysome thingsthatmight fix the problem?"She
says, "Yes."I type "paperjams" in the "problem"field and the right number
in the CCA field but nothinghappens.So I look up the code for paperjams on
my CCA paper cheat sheet. I finally get into CCA by using the right alpha-
numericcode for paperjams. I scan the list of questions that appearson the
screenandask the callerthe firstone: "Howmanycopies have you been doing
at a time?"since the instructionssay twenty-five is the limit. The caller says
people areonly allowed to do five at a time since the machineis for publicuse.
I readoff a suggestionfor adjustingthe feed trayand the next one which asks
if the papercould be damp. The caller responds,saying they have had very
dampweather,and I tell her to try a new reamof paperand recall thatI have
often heardSandy mentionthis to customers.There is anothersuggestion on
the list, but this seems like a good one, so I do not readthe final suggestion. I
do readsomethingaboutpressingthe startbuttontwice to the customer,but it
was not worded clearly and neither of us is quite sure of the step-by-step
process. Keeping the conversationnatural-soundingwhile wording the "Do
this,do that"commandsin a way thatdoes not soundcondescendingis difficult.
I tried to word my instructionsas I had heard other TeleReps do, in an "of
course you've done this" tone of voice, because the CCA instructionscome
from customermanualsand are mostly common knowledge for anyone who
has worked regularlywith copiers or electronic equipment.After telling the
caller to try changing the paperI tell her to call back and cancel the service
orderif thatworks andthankher for tryingthe steps as I have heardSandydo.
When I returnfrom lunch, the small screen on my phone console says, "4
calls waiting."The overheadLED screen which shows numberof customers
waiting and their averagetime of wait had displayedall O'searlier.Now it is
lit up with numberswhich keep escalating. A bell rings and everyone drops
whateverthey are doing to get on the phones and bring the numbersdown.
Everyone, including me, ignores CCA, turningall calls as much as possible
into vanilla service calls to get the backlog down as fast as possible.
The numbersarenot so severe now, my peakhadbeen 14 waiting,just after
I got on, now it shows only 6. I take two more calls but neitherwants to try
CCA. A third doesn't even know what area her jam is in: input, output,
whatever,so I try nothing. Customersdo not want to try things to fix their
machine very often. They would ratherjust call the technicianto fix it, even
whenjust changingthe paperor turningthe machineon andoff mightget them
backup andrunning.Gettingpeopleto attemptto fix theirown copieris another
level of TeleRep knowledge. The TeleRep has to read the customer's mood,
probetor contextualclues, such as whetherthe copier is nearbyand carefully
word the suggestions so thatthe customerdoes not feel thatthey arehaving to
do what their service contractshould provide.
During my two weeks on the phones, I remained dependent on my posted

documents for less frequently used codes but did acquire some tacit knowl-
edge. Doing the work myself, but dependent on the expertise of my experi-
enced neighbor, Sandy, I became particularly aware of her level of tacit
knowledge,includingnotjust themechanicsof takingservicecalls, including

work-arounds,but also her extensive knowledgeof all the variousmodels of
copiers and how to fix them over the phone. Some of this knowledge was
gained from her husband,who was a service technician,some from experi-
ence; but none of it was includedin companytrainingprograms.
By the end of the two weeks I had severalsuggestionsfor the reengineer-
ing design team about how to improvetraining.I also had suggestions for
changes to the antiquatedsystem which would makethe workpracticemore
efficient,as the threejobs becameintegratedinto one. However,sharingwhat
I learned with the reengineeringdesign team during our regular design
meetings was not to be. The structureof the design meetings focused on the
workat an abstractlevel andused a menu-drivenprocessthatwouldnot allow
the discussion of work details to take place. At the beginning, it certainly
appearedto our team of sociologists, concernedwith knowledge situatedin
practice,that having the workerswho actuallydid the work design the new
integratedwork process and its supportingtechnology was an excellent
insight on the partof management.Indeed,the worker-reengineers knew the
work in depth and in detail. However,whethermanagementintendedso or
not, the design practice itself and the documents for capturing it were
hamperedby management-drivenimperativesfor following a rigid process.
This ultimatelytook the formof a piece of softwarein which everythingwas
preordained,andthe designersonly filled in the blanks-another exampleof
buildingpolitics into artifacts.This, of course, workedagainstthe flexibility
crucial to innovative design work. Significantly, the introductionof this
software and the translationof the project into prespecified terminology
occurredwhen it lost its autonomyas a singularinitiative and was put, by
management,undera corporate-widereengineeringinitiative.At thatpoint,
which happenedto coincide with the arrivalof our team of sociologists, the
whole projectwas rewrittenin the largerinitiative'slanguage,the company's
versionof total qualitycontrol (TQC).
The "Quality Process": Quality or Control?
In contrastto the engineeringdesign world of negotiation,collaboration,

and/orpotentialconflict, negotiationover the documentsfor the redesignof
theactualworkhardlyever occurredat thereengineeringsite. Theexceptions
occurredwhen our research team raised questions about details of work
practice,althoughwe were often silenced for being "at the wrong level."6
The design process was constrainedboth by the design documentsand the
formatof the meetings.Managementperspectivesintervenedin the process,
not by their presence per se, but by the dictatedformatof design meetings
and design documents.Management-directedcorporateculturedictatedand
constrainedinteractionthroughits TQC conventions.These includedman-
dated step-by-step prescribedprocess for runningmeetings, documenting
what took place, and what action would follow. A menu-drivenprocess
formuladictatedstaying at a high level of abstractionbefore details of work
practicecould be considered.The formalismwas compoundedby the soft-
ware templates used to create visual documents. Such documents were
supposedto representworkprocessandwereconstantlyemployedas a means
of shorthandcommunicationin meetings.
Anyone making a presentationin the corporatesetting uses overhead
slides composed of boxes andbullets andpictograms.This computer-gener-
ated format has become the correct medium of corporatetalk; one is not
professional if one cannot reduce the message to sound bites, icons, and
bullets.7The requisiteexecutive summaryis intendedfor managerswho do
not have time to addressdetails. Concreteexamples of work practice that
supportsummarystatementsare left out. This detail is what broadergener-
alizations should be built on, so acceptanceof the broaderabstractionsis
dependenton trust that the person making these generalizationshas paid
attentionto the detail. However, all too often the abstractionscome down
from management'sgeneral impressions ratherthan being built up from
concretedata because the details of work practiceare associatedwith low-
status workers. A reengineeringteam must pay attention to the level of
practicaldetailthatcannotbe easily reducedto abstractions.The workhas to
be thoroughlyunderstoodbefore it can be redesigned.However, so much
time was spenton documentsat the high level (theirterm)of abstraction(my
addition)thatwhen attentionfinally turnedto detail,little or no time was left
because it had been eaten up in numerous two-day meetings devoted to
high-level discussions aroundhigh-level documents.
The structuredinteractionwas the result of the applicationof manage-
ment's interpretationof the so-called total quality process introduced by
W. EdwardsDeming via Japanesemarketsuccess.8This concept of quality
is based on the idea thatbusiness must anticipatethe needs of its customers,
transformthese needs into a useful and dependableproductthat meets or
exceeds customer expectations, and in so doing create a process that can
producetheproductatthe lowest possible cost so thatqualityandprofitability
become synonymous.Deming recommendeda cyclical perspectivepromot-
ing ongoing change and improvement,not only of the productproduced,but
also of the structureand processes of the firm. Cooperation among all
departmentswas intendedto facilitateredefinitionof the customer's needs
in a changing environment.Although corporateadaptationsof Deming's
ideas vary,any good idea can always be reified andformula-izedas it moves

from originatorto management-mandated applications.Deming also advo-
cated commitmentto long-termstrategies;eliminationof numericalgoals
and quotas that focus managementattention on ends ratherthan means;
driving out fear so thatpeople can work effectively; leadershipthat guides
ratherthan polices; collaborationwith suppliersratherthancontractsbased
only on lowest price;employee duties thatare meaningful,accompaniedby
programsfor educationand self-improvement;eliminationof mass inspec-
tion, emphasison teamworkinsteadof individualismandcompetitionamong
employees, includingon-the-jobtrainingin which managersparticipate;and
eliminationof the individualmeritsystem and the numericmeasurementof
job productivity.Deming maintainedthat if a system is functioning with
quality processes in place, it will not generatewaste or faulty products.He
also advocated the eradicationof slogans and exhortationsbecause they
imply thatreformlies with the employeesratherthanin a functioningsystem.
According to Deming, change must be holistic, encompassingevery aspect
of the organization.The structureto facilitatesuchtransformationmustcome
from top management(Townsendand Gebhardt1992).
Deming's TQC process recommendationsmay or may not have been the
primaryreasonsfor Japaneseascendancein a competitivemarket.American
business has incorporatedDeming's perspective in many different ways,
some more successful than others. In many cases, managementhas taken
some ideas and not others, abstracting,reinterpreting,and reifying them,
using some of the very perspectivesand processes of which Deming was
critical.Too often, management-mandated workshopsthroughoutcorporate
Americahave institutionalizedslogans, acronyms,and variousprocess for-
mulas and checklists promotingquality processes and attemptsto measure
quality. Contraryto Taylor's original ideas, just as scientific management
came to dictateto laborthe one rightway thatwork should be done, so also
managementhas all too often selected some, but not all, of Deming's ideas
and applied them piecemeal from a top-down perspectivethat missed the
holistic characterof the nonhierarchicalteamworkDeming advocated.
This was the case in the companywheremy researchwas conducted.The
TQC process, initiated in the early 1980s, was conceived as a top-down
corporateculturechange.Eachlevel of managementwas supposedto master
an abstractedand truncatedversion of Deming's ideas, "cascadingit down-
ward"in companyjargon. TQC was handeddown the managementchain,
startingeach time with managersandtheirstaff,who thentaughtit to the next
level, downward, and eventually to employees. Pace and Argona (1991)
reportthe conflict between this top-downperspectiveand a joint union and
managementprojectto improvethe qualityof work life (QWL) in the same
Objects 157
/ Roleof Material
Henderson
company.They also point out, to the company'scredit,thatit allowed QWL

representativesto modifythe TQCcurriculumanddeliverit to the workforce.
In the past decade, furthermodificationsof Deming's concepts have taken
place.9The companyhas won nationalqualityawardsand has implemented
Deming's ideas such as employee self-improvementprogramsand an em-
phasis on teamwork.And yet it maintainspracticesantitheticalto Deming's
TQC. The company still uses quantitativemeasures for everything from
numberof minutes spent per customertelephonecall to the setting of sales
goals. It still maintainsindividuallybasedmeritevaluationanda proliferation
of empty slogans that promote the idea that delivering quality is the sole
responsibilityof the worker.The following analysis traces this company's
version of its TQC formula in the structureand language of the design
documentsprovidedto the reengineersin softwaretemplates.
Ourteamjoined the NationalDesign ProgramTeamafterits membershad
completed the initial conceptual stage of the reengineeringproject. The
design meetings we attendedbegan with large stacks of paper documents
being passedout to all present.Sometimesthese were blackandwhite copies
of multicoloredslides. Othertimes they were sets of paperdocumentsalone.
The personwho was responsiblefor the TQC action item for a subjectarea
would presentthe documentby readingthroughit. Others,whose specialty
area was the same, would respond. There was no sense of innovation or
creativity.Therewas no playing with the designedobjectas describedby the
engineers who designed the surgical instrument.There was no room to
explore related areas. The contents of the document set the agenda for
discussion, and little deviationwas made from it. The documentsserved as
a checklist for things alreadyknown. Representationsof work stayed at a
very abstractlevel. No unrepresentedtacit knowledge was elicited. If more
concretesuggestions or observationswere made, the speakerwas often told
they were at the wronglevel. The team was said to be workingat a very high
level, butthatmeanta very low level of understandingof the actualwork.As
participantsengaged in documentingthe work practicethatwas being reen-
gineered, our team members attemptedto make comments about actual
practice. But we also were told that we were at the wrong level. In other
words, our input was too concrete in a discussion where things were to be
kept very general and abstract.After being repeatedly reprimanded,all
participants,includingourresearchteam,learnednot to mentionsuchdetails.
We awaitedthe appropriatetime to raise detailedwork practiceissues.
As time passed, the National Design ProgramTeam worked from high
into medium levels. When flowchartswere finally broughtout to discuss
process, it appearedwe were finally at a level to addresswork practice.But
the company'sand hence the design team'sdefinitionof process was not the
same as practice.Process was formalizedand prescribedso thateven here it

was hardto introduceattentionto detail. Time was short,the pilot initiation
deadlinewas approaching,andmanyothertopics hadto be covered.So much
time had been spenton documentsat a high level thatwhen the time came to
addressthe details of practice,therewas very little time left to do so.
The "Design Documents"
A series of documents illustrates the company's and the design program

team's notion of moving from a very high level to a detail level. The
representationof the business profile of the overalljob integration-planis a
good example of a very high level document(see Figure 1) composed for
broad dissemination at the managementlevel with figures expressed in
millions andbillions.i?
The second example (Figure 2) is a working document used by the
programdesign team. It presentsthe componentpartsof the new integrated
job, dividedinto blocks thatrepresentactualjob descriptions.This is a design
process document, a detailed representationof the actual processes to be
incorporatedinto the new integratedjob, designatedby numbersthat index
official writtenjob descriptions.At the next level down (Figure 3), each
process is furtherdetailed. A designated specialist who had actually per-
formed the job being analyzed worked on the document to be sure that it
includedall the pertinentinformationandthenpresentedit backto the design
team for furtherreview. Feedbackgenerally came from those on the team
who had the samejob expertise.
The thirddocument(Figure3) shows how one of the business subproc-
esses of technicalservice management(9.0), the so-called service call man-
agement(9.3), can be brokendownfurtherinto two subsubprocesses:service
call handling(9.3.1.1) and service dispatch(9.3.1.2). Althoughthe number-
ing system seems to imply thatwe aremoving into a very detailed area,the
job practicesare abstractedinto the preformulatedjargon of the company's
dictated process abstractions:objective, process owner, inputs, outputs,
documents.To make sense of this language use, one must know the com-
pany's specializedcodes andhave some experiencewith thejob itself. In this
respect, the documentresemblesthe design documentsin mechanicalengi-
neering that are intended for specialty engineers. The electrical engineer
contributesone set of codes to a particularsegmentof the design documents
while the fluids engineerandthe structuralengineerconstructothersegments
using theirparticularsets of symbols. Despite the fact thatthe reengineering
documentuses English, its languageis neitheruniversalnor transparent.In
|Integrated Customer Service
BUSINESS PROFILE |
|BUSINESS | | CUSTOMER CARE, PROBLEM RESOLUTION, TELEMARKETING, SUPPL

PROCESS | CUSTOMER SERVICE
r SUPPLY ^ r TELE- ? r CUSTOMER '

SUB-
MARKETING MARKETING SERVICE
PROCESS l l SUPPORT
-$ ORDERS I S N REVENUE CALLS FROM

- ^WERS.
~AND
GROWING 'UBUD
^^^^^H ~- LEADS CLST
f l GENERATED CUSTOMSET
- CALIJMAIL 11
W
WITH
CONTACTS T \^T - ^
CUSTOMERS I
- CUSTOMER - CUSTOMER
CNAT CONTACTS
KEY MESSAGE:
* There are million customer contacts done by today's functionally oriented personnel,
processes and technologies.
* Each of these customer contacts represents a 'Moment of Truth' to satisfy customers an
of Business Priorities.
Figure 1
Process
Sub-
Process
Figure 2
SERVICEMANAG
USCOBUSINESSAREA:TECHNICAL
KEYBUSINESSPROCESS:SERVICECALLMANAGE
KEYWORKPROCESS:SERVICEDELIVERY(9.
9.3.1.1
&|k SERVICE CALLHANDLING . E SERVICEDISP
Providefront.endproductsupport to customers through

accurate rvice entlement Identification,problem Ensure timelydispatch and hand
Objectve: Identiflca Remon CallAshtance, and prompt analystfserviceengineer.
ecalation to *nelylthwrvlceengineer.
Process WorldwideCustomerServices a
O w n erCustomerServices Support Support
*Service Requests *Pending Ser ice Calls

mpins. *OutboundSeviCe Alerts *PendingSoftwareCalls
*HardwaceSertice Callneed Identified

*Softwne,Service Callneed Identified *.HrdwareService Calldispatc
Outputs: *Co atlon Callscheduled *Soltwar Service Callhanded-
*AccountProfileupdatd
*RemoteCallAssistance successfully completed
*Service CaJlRecord *Service CallRecord

.*SofwareCallRecord *SoftwareCallRecord
Documents Remot CallA$$1stlnceSolve Record *RemoteCallAssistance Solve
*Accountprofileupdates *Accountprofileupdates
Figure 3
WORK PROCESS: SERVICECALL HANDLING9.3.1.1

Provide Front-endproduct support to customers through accurate service
Objective: entitlement Identification,problem identification, Remote CallAssistance, and
promptescalation to analystlservice engineer.
Process Owner: Customer Services Support
Inputs* Service Requests

Outbound Service Alerts
Key ProcessSteps: >^ \
Confimatio
s cheduld
A t proflC
Remote a
Dttermloe *ny i^^^^ Csil AsiistaBce\
\, reuiirmenti sSBQS\ Dispstch procesi
Oupt: Hardware Service call need idenftifed, Software call need Ideneified,
pUS ConfirmnationcaSIscheduled, Account profile update, Remote aill
Assistance successfully completed
C ustomer , Field Technicians, Analysts
Service Call Record,Software Call Record,

Documents: Remote CallA.istance Solve Record, Account
Profile updates
Benefits:
Quality
Cost - Reduced on4ite service calls
Del'verv - Customer downtimereduced
Figure4
this document, the representation of the detail is constrained by the inflexi-

bility of the company's formalized interpretation of process components.
This interpretation dictates what can be expressed and prohibits descriptions
of actual practice.
The next level document (Figure 4) represents the redesigned job, but it
does not take us much closer to practice either. Three quarters of it is identical
to the document in Figure 3. What has been added are the customer, benefits,
and a visual representationof key process steps. A practice-orientedaction

researchertryingto participatein this design process inevitablyexperiences
frustration.Every time my colleagues and I tried to clarify practice or
contributeinformationon knowledge in practice,we were told we were at
the wrong level. This document can be read as an actual picture of our
frustration:Each circle in the key process steps identifies the sequence of
events the TeleRepgoes throughin takingandplacing a service call request.
However,the circles literallycircumscribethe spaceavailablefor the descrip-
tion of steps,makingit very clearthatthis is still the wronglevel for capturing
actualpractice.Furtherproof of how little design manipulationoccurredat
these levels is providedby the comparisonof the originaldocumentfrom 25
January1995 (Figure4) with ones from 8 February1995 (Figure5) and 16
February1995 (Figure6). The only differencebetweenthe first two of these
documents is that process owner has been changed, the quality line under
benefits has been filled in, and the borderlines have been removed. In the
next iteration,the only change is computer-generatedaesthetics, the docu-
ment has been moved to landscapeconfiguration,and the backgroundshad-
ing has been removed.
But I still hoped that, at the most detailed level, the process flow chart
(Figure 7) representingthe job as redesigned would finally address the
contingencies of the work itself. Time was runningout. Plans for the pilot
were being made anddesign teammemberswere being called away to other
responsibilities.On the day when we finally discussed things at the level of
work practices,only some of the National Design ProgramTeammembers
were presentand the projectmanagerwas not there.The flowchartthatwas
passed around,dated7 February,was no differentfrom its predecessorfrom
25 January.Fortunately,those presentdid include areaexpertswith service-
call experience.When I startedaskingquestionsaboutthe actualwork, they
listened andresponded.Ourinteractionsgeneratedchanges to the chartthat
arehandwrittenin Figure8. At least some of the messiness and ambiguityof
the work got discussed, although even there, little was captured in the
documents.The whole sequenceof eventsthatoccurredwhen a TeleRepused
remotecall assistance(RCA),11thatrequiredknowledge of how to coax the
caller to turn their machine off and on or run some copies, as well as
familiaritywith machinetypes and common problems,has been reducedto
a flowchart decision. Although the flowchartcould be used as a centering
device for discussion of the detailsof this workthatcould be improvedin the
redesign, time and energy for examinationof how the work really got done
had run out. Had our researchteam not interruptedthe structureddesign
sequence earlier,with our own workshop, our suggestions relevant to the
details of the work would hardlyhave been discussed.
XVORKPROCESS: SERVICECALL HANDLING9.3.1.1
Provide Front-end product support to customers through accuraet service

Objective: entitlement identification, problem identification, Remote Call Assistance, and
prompt escalation to analyst/service engineer.
Process Owner: 8erviceCallHandling

ProcessOwner
Inputs: ServiceRequests
Outbound Service Alerts
Key Process Steps:

J Cutomer\ t k,
| Account
*nd 5h|_
& PrblemB8iB_
sai InormtbnA t _SS rob
01
Mianose
Proble
4Perform 4. Remote'
Detennine.ny CallAsi,stnce
follow-up ~fl (exittoService
\ requlremenU/ % Dlispatch
process
/ fedd
Harowarcservice call need Idenfified, Software call need identified,

Outputs: Confirmation cell scheduled, Account profilc updated. Remote Call
Assistance successfully completed
Customer: Customer, Field Technicians, Analysts
nArnm-4o Service Call Record, 8offwarn Call Record,

Remote Call Assistance Solve Record, Account
Profile uDdates
Benefits:
Quality ,Accuracy of Informationgiven to customer (.e., Cuebase Info.)
Cost -Cost per telephone Call
Dellewv - Response to customer phone call
Figure5
Because of this change of communication format, some innovations based

on attention to practice were made in the pilot. Among these were (1)
awareness of the importance of at-a-glance support documents so that some
paper documents, already in use, were captured and expanded; (2) awareness
of the significance of TeleRep interaction and sharing of information and,
hence, the importance of cubicle wall height for both easy communication
between work groups and some level of sound buffer, also applied in the pilot;
WORKPROCESS:SERVICECALLHANDLING9.3.1.1
| Ob
tii* i ve:
Ujec
!
|
ProvideFront-endproductsupportto customers through accurate service
| problemidentification,Remote CallAssistance, and promptescalation to
Process Owner: t&ece c.ll andllng Proc

Ownj
Inputs'
-npu8: ServiceRequests|
11~I
I1 OutboundService Alerts
KeyProcessSteps: , !
y Obt.l Cu tomer.
^ -'^l Account ,nd )
f Verifycustomer ~
ndclose
sUtisfctloni ] [
\^ nicrvcall y \
> \\ <r 4 > Perform R

Detetme ny 1 ff Aufista
%fo.ljw-up / . Selrice
'^ reqnlrement p^ sjroces I
II Outputs: * HardwareService call need identified,Softwarecall need identified, Confir

I t Account profileupdated, Remote CallAssistance successfully completed
| Customer: | ^ustomer, FieldTechnicians,Analysts
se* Record,SoftwareCallRecord, Remote Call

Ince CallSolve
I)ocumenc. | Assistance Record, Account Profileupdates
Benefits:
QuarAy - Accuracyof informationgiven to customer (i.e., Casebase info.)
Cost - Cost per telephone Call
Delivery :Response to customer phone call
Figure 6
3
U1
?
s
3
o~X O:$
oE f
O)
'0
CfJ
0:- . 0
!
0
I
QC.
]
3 s;
0
'
-
o ..
s
UZ
- K
E e
_ a
- j
gE3
' 8 X w
rI
2~
166
EU
I
L)
'- -
I
oo
o X
)
L)
3?
X
o:
2C
o
;
?l}
*..C
I i
on s t
i
180
5
l00
*o,
I
167
168 & HumanValues
Science,Technology,
(3) awarenessthatthe"smartsystem"for solving machineproblemsover the

phone was not being used in practiceas it was designed to be, which was
addressedin existing practiceby soliciting morefeedback fromusers andby
a company-sponsored,month-longcompetitionamong users for the highest
numberof solves (solutions).The updatedsystem was used in the pilot.
Issues thatwere not addressedin design meetingsbecause the discussion
of actualpracticewas putoff for so long includedthe omission of knowledge
in practiceresidingin the sequenceof customer-TeleRepinteractionsduring
the use of RCA. In the flowchartmodel of customer-TeleRepinteraction,the
juncture at which the TeleRepinitiatedsolving the machineproblemon the
phone is shown as a simple yes/no decision tree.However, in practice,much
skill in handlingthe customerwas necessary to get them even to try to fix
their own machine;in fact, many refused. Successful skills involved using
the right language and tone along with holding out the potential that the
customer's copier could be back up and runningright away. Another tech-
nique, useful if the caller's telephone was remote from the copier, was
informingthe customerthatthe service call was being placed, as usual, and
asking them to call back and cancel if the recommendedsteps solved the
problem. In all cases, judgment of the customer's mood, their kind of
business, along with the level of difficulty of the problemrequiredas much
skill and knowledge as knowing the various machinemodels, their typical
problems, and the right recommendationsfor solving those problems,
whetherfound in the databaseand read from the screen in a mannerto keep
conversationflowing or recitedfromthe TeleRep'smemory.Such skills and
knowledge were neitheracknowledgedby managementnor taughtin train-
ing. The assumptionwas that TeleReps knew little and were dependenton
the smartsystem to accomplishsuch solves.12Discussion of how TeleReps
actually solve machine problemsover the phone would have revealed the
mistakennatureof such assumptionsand would have shown the shortcom-
ings of the smartsystem. Had this taken place early in the design process,
less emphasis could have been put into trying to make the ill-suited, case-
basedsmartsystemwork,andmoreemphasiscouldhave beenputon training
TeleReps in machine-problem-solvingskills. Our IRL team's participation
in the design and constructionof trainingcurriculumbroughtemphasis on
learningthroughpracticewith expertpeers as tutors,facilitatingsome tacit
knowledge transfer.However, the mind-set that credited the technology
ratherthanthe workerdid not change.It mighthave at least been challenged
had such practice-orienteddiscussion takenplace in the design meetings.
Anotherexampleof the resultsof ignoringpracticewas the destructionof
useful resources because they were put on-line with no awarenessof their
use in practice.The inclinationto have technicianssimply put paperdocu-
ments on-line withoutunderstandinghow they are actuallyused in practice

is a common pitfall in the use of computersystems. The particularexample
here applies to supplies sales ratherthan service calls. Indiscriminatedigi-
talizationof existing paperresourceshadresultedin the destructionof useful
documentssuch as the one used for floor (lowest allowed) prices in supplies
sales. As a paperdocument,referredto as the hible, it had been individually
customizedwith highlightingandbook-markingfor often-usedsections, and
handwrittenmarginnotes for specialcases andupdates.Althoughinteractive
technology was capable of incorporatingsuch functions,the documentwas
put on-line in a form thatmadeit impossiblefor people to customizeto their
use, andit became a deaddocument.As a compilationof unuseddata,it was
a waste of technology and of a previouslyvaluableresource.We were able
to addressthis issue in memos fromthe IRLteamduringthe pilot period,but
it would have been preferable,however, if the importanceof designing
on-line documentsto emulatetheiruse in practicewere discussed earlierin
the process with the whole team of designer-users.Althoughsome attention
was given laterto which documentsshouldbe put on-line, when the techni-
cians responsible for this task requestedthat our team give them a list of
documentsto be digitalized,they expected that we would provide a yes/no
type of list ratherthan an answer demonstratingawareness that how the
informationcontainedin the paperdocumentwas used in practicemattered
for the format in which it should be made available electronically.When
earlieruse practiceis not emulated,the assets of the original-such as speed
of access facilitated by at-a-glance structureand location, and its custo-
mizable flexibility-are lost.
Conclusion
The comparisonof the two design cultures-industrial designengineering

practice and this particularreengineeringpractice-reveals a contrast in
termsof flexibility and innovation.
The mechanicaldesign engineerssketchedto one anotheron scrappaper
and in the marginsof official drawingsto convey meaning.The reengineers
were boundby predefinedtemplatesand icons in theircomputerprogram.
Industrialdesign engineers did not follow design school protocols in
practice,preferringto work on differentlevels and differentsegments of the
design simultaneously.The reengineershad to work using dictated steps,
controlledby management-mandated structureof theTQCprocess,deviation
from which was quickly pointedout. The work was a lock-step hierarchical
process that stayed at a high level of abstractionto the extent thatlittle time
was left for the crucial details of work practice that could make or break the
project's effectiveness.
The industrial design engineers used mixed electronic world and paper
world practices to create visual representations, choosing the format most
appropriate for a given context. The reengineers used only the computer
software templates and, given the conventions of presentation and the de-
mands of management, had no other choices.
Corporate management has created a myth that it is handing over crucial
decisions to empowered work groups, stepping aside so that workers can
redesign their own work practice and its technical support. However, corpo-
rate programs to ensure quality through uniformity of practice also ensure
that little innovation takes place. To its credit, this is not what management
is trying to do. It is trying to give guidance and structure. However, just like
Taylor's ideas that have come down to us through scientific management,
what might have been a good idea to start with in one particular application,
when reified and made into a formula, results in the loss of the creative
flexibility and of the innovative informal processes that are the cornerstones
of doing things well. The real dangers are less the software packages that
create concepts and visual representations of work than the assumptions that
one idea will work for every application and that the formalization of good
ideas will produce standardized good work. Such reifications cause the loss
of the necessary messy, informal detail and patched-together character of real
work and knowledge in practice that made them fit the situation in the first
place. Streamlined and abstracted knowledge in executive summaries and
prepackaged formats may serve the purposes of review for busy executives,
but such standardization can result in slippage and misfit between plans,
appropriate action, and anticipated outcomes.
Notes
1. See,forinstance,Rudwick(1976,1992),Mukerji (1984),Lynch(1985a,1985b,1988),
LatourandWoolgar(1979),LynchandWoolgar(1988a,1988b),AmannandKnorrCetina
(1988),andKnorrCetina(1990).
2. SeeAllwoodandKalen(1994)forareviewof theliteratureonthelimitations of computer
assisteddesign(CAD).
3. StarandGriesemer (1989,393)defineboundary objectsas "objectsthatarebothplastic
enoughto adaptto localneedsandconstraintsof theseveralpartiesemployingthem,yetrobust
enoughto maintain a commonidentityacrosssites."
4. Thiswasa teamprojectin actionresearchsponsored throughtheInstituteforResearch
on Learning(IRL)and fundedby the unnamedcorporation at whose site the studywas
conducted.
andcompanynamesarepseudonyms.
5. All individual
6. Once the members of our team realized we would not be able to convey the crucial
informationwe needed to get across to the design team because of the structureof the design
meeting format,we introduceda differentinteractionalstructurein the form of a workshopto
presentour findings. However, this was an exception introducedby our presence, and was not
partof the usual process.
7. The use of slick corporatepresentationslides is similar to the art history example of
Alberti(1404-1472), who had to presenthis argumentsregardingthe importanceof perspective
in paintingin Ciceronianrhetoric,regardlessof its inappropriatenessfor the subject,to be taken
seriouslyby the Medici court,because such rhetoricwas the respectedmode of discourseof the
day. Overheadslides presentingmaterialin boxes, icons, and bullets is the Ciceronianrhetoric
of today's corporateculture.
8. Quality became the corporatefetish of the 1980s and 1990s, anotherby-productof
shrinkingsingle company marketshares due to increasedinternationalcompetition.In many
sectors, the focus has been deemed the new battlegroundfor corporatesurvivaland prosperity.
Deming, originatorof the conceptand a U.S. citizen, founda receptiveaudiencefor his quality
philosophyin the desperateeconomy of postwarJapan,his ideas having been rejectedby the
then-booming U.S. business community. Declines in U.S. market shares due to Japanese
competitionin the 1970s broughtrenewedattentionto Deming's ideas.
9. Marcia Porterfield,"The Quality Philosophy and One Corporation'sInterpretation"
(unpublishedmanuscript,n.d.).
10. This representationalso includestelemarketingthatwas to be incorporated,althoughthe
National Design ProgramTeam was working on integrationof supply marketing,customer
service support,and customeradministration.
11.This is an updatedtermto includenew supportsystemsto help customersfix theircopiers
with database-supportedadvice from TeleReps.CCA, the older system, is discussed earlierin
the field notes.
12. How such solves were actuallyaccomplishedis discussed in Vinkhuyzenand Whalen
(1996) and Whalenand Henderson(1996).
References
Allwood, Martin,andThomasKalen.1994. Usabilityin CAD: A psychologicalperspective.The

InternationalJournalof HumanFactors in Manufacturing4 (2): 145-65.
Alpers, Svetlana. 1983. The art of describing.Chicago:Universityof Chicago Press.
Amann, Klaus, and KarinKnorrCetina. 1988. The fixationof visual evidence. HumanStudies
11:133-69.
Arnheim,Rudolf. 1969. Visualthinking.Berkeley:Universityof CaliforniaPress.
. 1993. Sketchingand the psychology of design. Design Issues 9:15-19.
Bijker,Wiebe. 1995. Of bicycles, bakelites,and bulbs. Cambridge,MA: MIT Press.
Bly, Sara. 1988. A use of drawingsurfacesin differentcollaborativesettings.Paperpresentedat
the Second Conference on Computer-SupportedCooperative Work, Portland, Oregon,
September.
Bowker,Geof. 1988. Picturesfrom the subsoil, 1939. In Picturingpower,visual depictionsand
social relations,editedby GordonFyfe andJohnLaw,221-54. London:Routledge& Kegan
Paul.
Bucciarelli,Louis L. 1994. Designing engineers.Cambridge,MA: MIT Press.
Buchanan,Richard,and VictorMargolin,eds. 1995. Discovering design. Chicago:University

of ChicagoPress.
Callon, Michel. 1986a. The sociology of an actor-network:The case of the electric vehicle. In
Mappingthe dynamicsof science and technology,edited by Michel Callon,John Law, and
Arie Rip, 19-34. Basingstoke,UK: Macmillan.
. 1986b. Some elementsof a sociology of translation:Domesticationof the scallops and
the fishermenof St. BrieuxBay. In Power,action and belief:A new sociology of knowledge?
edited by John Law, 196-229. London:Routledge& Kegan Paul.
Collins, HarryM. 1974. The TEA set: Tacitknowledgeandscientific networks.Science Studies
4:165-86.
Downey, Gary. 1992a. CAD/CAMsaves the nation?Knowledgeand Society9:143-68.
. 1992b. Humanagency in CAD/CAMtechnology.AnthropologyToday8:2-6.
. 1995. Steeringtechnologydevelopmentthroughcomputer-aideddesign. In Managing
technologyin society, editedby A. Rip, T. Misa, and T. Schot, 83-110. London:Pinter.
Ferguson,Eugene. 1976. The mind'seye: Nonverbalthoughtin technology.Science 197:827.
. 1992. Engineeringand the mind'seye. Cambridge,MA: MIT Press.
Fujimura,Joan. 1988. The molecular biology bandwagonin cancer research:Where social
worlds meet. Social Problems35:261-83.
Fyfe, Gordon,andJohnLaw,eds. 1988. Picturingpower: Visualdepictionsand social relations.
London:Routledge& KeganPaul.
Gorman, Michael, and BernardCarlson. 1993. Invention as re-representation:The role of
sketchesin the developmentof the telephone.Paperpresentedat the PrincetonWorkshopon
Cognitive Approachesto Visualizationin Science and Technology,February.
Harper,R. R., J. A. Hughes,and D. Z. Shapiro.1991. Workingin harmony:An examinationof
computertechnology in air traffic control. In Studies in computersupportedcooperative
work: Theorypractice and design, edited by J. M. Bowers and S. D. Benford, 225-34.
Amsterdam:Elsevier.
Henderson,Kathryn.1991. Flexible sketches and inflexible databases. Science, Technology,&
HumanValues16:448-73.
. 1993. Mixed practicesin the visual cultureof design engineering.Paperpresentedat
the PadisWorkshopon Social Science Research,TechnicalSystems and CooperativeWork,
April.
. 1994. How and why changes from one visual cultureto anothercause restructuringin
design engineering.Paperpresentedat the Conferenceon VisualRepresentationsin Scien-
tific Practice,Galveston,Texas,April.
. 1995a. The politicalcareerof a prototype:Visualrepresentationin design engineering.
Social Problems42:274-99.
. 1995b.The visual cultureof engineers.In Culturesof computing,editedby Leigh Star,
196-218. Oxford,UK: Blackwell.
. 1998. On line and on paper: Visual representations,visual culture, and computer
graphicsin design engineering.Cambridge,MA: MIT Press.
Hindle,B. 1981. Emulationand invention.New York:W. W. Norton.
Hutchins,Edwin. 1991. The social organizationof distributedcognition. In Perspectives on
socially sharedcognition,editedby LaurenResnickandJohnLevine, 283-304. Washington,
DC: APA.
. 1995. Cognitionin the wild. Cambridge,MA: MIT Press.
KnorrCetina,Karin. 1990. Imagedissection in naturalscientific inquiry.Science, Technology,
& HumanValues15:259-83.
Latour,Bruno. 1986. Visualizationand cognition:Thinkingwith eyes and hands. Knowledge

and Society: Studiesin the Sociology of CulturePast and Present6:1-40.
. 1987. Science in action: How to follow scientists and engineers through society.
Cambridge,MA: HarvardUniversityPress.
. 1992. Where are the missing masses? The sociology of a few mundaneartifacts.In
Shapingtechnology,buildingsociety, edited by W. Bijkerand J. Law, 225-58. Cambridge,
MA: MIT Press.
Latour,Bruno, and Steve Woolgar.1979. Laboratorylife: The constructionof scientificfacts.
Beverly Hills, CA: Sage.
Lave, Jean. 1988. Cognitionin practice. Cambridge,MA: CambridgeUniversityPress.
Law, John. 1987a. On the social explanationof technicalchange: The case of the Portuguese
maritimeexpansion.Technologyand Culture28:227-52.
. 1987b. Technology and heterogeneousengineering:The case of Portugueseexpan-
sion. In The social constructionof technological systems, edited by W. Bijker,T. Hughes,
and T. Pinch, 111-34. Cambridge,MA: MIT Press.
Luff, P., C. Heath, and D. Greatbatch.1992. Tasks-in-interaction:Paper and screen based
documentationin collaborativeactivity.In Proceedings,10thInterdisciplinaryWorkshopon
"Informaticsin Psychology": CognitiveAspectsof VisualLanguageand VisualInterfaces,
Sharding,Australia,May 1991.
Lynch,Michael. 1985a. Art and artifactin laboratoryscience. London,Beverly Hills: Sage.
. 1985b. Discipline and the materialform of images: Analysis of scientific visibility.
Social Studiesof Science 15:37-66.
. 1988. The externalizedretina:Selection and mathematizationin the visual documenta-
tion of objects in the life sciences. HumanStudies 11:201-34.
Lynch, Michael, and Steve Woolgar,eds. 1988a. Introduction:Sociological orientationsto
representationalpracticein science. HumanStudies 11:99-116.
. 1988b. Representationin scientific practice.Special issue on Representationin Scien-
tific Practice.HumanStudies 11 (2-3).
Margolin, Victor. 1989. Introduction.In Design discourse, edited by Victor Margolin, 3-28.
Chicago:Universityof Chicago Press.
Margolin,Victor,and RichardBuchanan,eds. 1995. The idea of design. Cambridge,MA: MIT
Press.
Mukerji,Chandra.1984. Visual language in science and the exercise of power: The case of
cartographyin early modem Europe.Studiesin VisualAnthropology10:30-45.
Pace, Larry,and DominickArgona. 1991. Participatoryactionresearch.In Participatoryaction
research,edited by WilliamFoote Whyte, 56-69. NewburyPark,CA: Sage.
Petroski,Henry.1985. Toengineer is human.New York:St. Martin's.
Polanyi, Michael. 1958. Personal knowledge: Towardsa post-critical philosophy. London:
Routledge& Kegan Paul.
. 1967. The tacit dimension.London:Routledge& KeganPaul.
Rudwick,Martin.1976. The emergenceof a visual languagefor geological science 1760-1840.
Historyof Science 14:148-95.
. 1992. Scenesfrom deep time. Chicago:Universityof Chicago Press.
Star,SusanLeigh. 1989. The structureof ill-structuredsolutions:Heterogeneousproblem-solv-
ing, boundaryobjects and distributedartificialintelligence. In Distributedartificial intelli-
gence 2, edited by Michael Huhns and Les Gasser, 37-54. Menlo Park, CA: Morgan
Kauffman.
Star,Susan L., and James R. Griesemer.1989. Institutionalecology, "translations,"and coher-

ence: Amateursandprofessionalsin Berkeley'sMuseumof VertebrateZoology 1907-1939.
Social Studiesof Science 19:387-420.
Suchman,Lucy. 1987. Plans and situatedactions: The problemof human-machinecommuni-
cation. Cambridge,MA: CambridgeUniversityPress.
Tang,J., and L. Leifer. 1988. A frameworkfor understandingthe workspaceactivity of design
teams. In Computer-supported cooperativework. San Mateo, CA: MorganKauffman.
Townsend,Patrick,andJoanGebhardt.1992. Qualityin action. New York:JohnWiley & Sons.
Turnbull,David. 1993. The ad hoc collective workof buildingGothiccathedralswithtemplates,
string,and geometry.Science, Technology,& HumanValues18:315-40.
Ullman,D., L. Stauffer,and T.Dietterich.1987. TowardexpertCAD. Computersin Mechanical
Engineering 16:56-63.
Vincenti,WalterG. 1990. Whatengineers know and how they know it. Baltimore,MD: Johns
HopkinsUniversityPress.
Vinkhuyzen,Erik,andJackWhalen. 1996.Machineintelligenceandhumaninteraction:Solving
problemswith documentmachines over the telephone.Paperpresentedat the meetings of
the AmericanSociological Association,New York,August.
Vitta,Maurizio. 1989. The meaningof design. In Design discourse, editedby VictorMargolin,
31-41. Chicago:Universityof ChicagoPress.
Whalen, J., K. Henderson,and M. Whalen. 1995. On line, on paper,and on the phone: The
choreographyof work. Paper presented at the meetings of the American Sociological
Association,Washington,DC, August.
Whalen, Marilyn,and KathrynHenderson.1996. Outstandingsmartsystems. Paperpresented
at the meetings of the AmericanSociological Association,New York,August.
Winner,Langdon. 1985. Do artifactshave politics? In The social shaping,of technology,edited
by Donald MacKenzieand Judy Wajcman,26-38. Philadelphia:Open UniversityPress.
. 1995. Political ergonomics. In Discovering design, edited by R. Buchanan and
V. Margolin, 146-70. Chicago:Universityof Chicago Press.
Kathryn Henderson earned her Ph.D. in sociology and M.EA. in art criticism at
Universityof California,San Diego. She is currentlyan assistantprofe.vssor
in sociology
at TexasA&M University(College Station,TX77843-4351). Her latest book is On Line
andon Paper:VisualRepresentations,VisualCulture,andComputerGraphicsin Design
Engineering(MITPress, 1998). She is currentlywritingaboutthe use ofpaper documents
in computer-supportedwork and researching the actor networks of the grassroots
sustainablebuildingmovement.

Material Objects in Design Process

Загружено:

Сведения о документе

Исходное описание:

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Material Objects in Design Process

Загружено:

Авторское право:

Доступные форматы

!"#$%&'#$&($)*+#,-*'$./0#1+2$-3$+"#$4#2-53$6,&1#227$8$9&:;*,-2&3$&($!

This article comparestwo culturesof engineeringdesign, one flexible and interactive,

This articlecomparestwo culturesof engineeringdesign, one flexible and

Science, Technology,& HumanValues,Vol. 23 No. 2, Spring 1998 139-174

influencingthe rangeof creativityandinnovationin a given sociotechnologi-

studiesby engineersthemselves(Bucciarelli1994) also documentthe impor-

The contingenciesof automationin work worlds thathad been tradition-

Visual Documents and Situated

technologies but also in the constructionof shared and individual work

Whenyou do draftingby hand,a lot of the thoughtprocesscomes when you're

Thus, as individualthinkingtools, sketches help to captureconcepts in a

Duringa study at a high-precisionmedical optics firm I observedeven more

a sketchwith one implement,each alternativelytakingthe pen fromthe hand

Whenyou're doing the mechanicaldesign work you have to be very awareof

Because he knows from experiencethatdrawingscan be misinterpreted

It [a drawing]helps in termsof, "Ohit looks pretty"or "Howdoes this [work],"

I'm basicallythe personwho comes in and says, "Ah,this isn't workingright"

Such play with the prototypewas greatlyadvancedby the engagementof

Reengineering a Call Center

The site of this studywas one of severalcall centersthatprovidessupplies

Headphones connect her or him to the computerizedtelephone system,

I plugmy headset into thephoneconsole in my assignedcubicle,next to Sandy,

rightcomer of the screen,which I also missed in my visual scan as I readdown

The next call is a legitimateENOF [equipmentnot on file]. The serialnumber

One of the forms of tacit knowledge I had to develop as a TeleRep was

I get call from an elementaryschool thathas "donot service, no renewalPC,"

Although I had been through training, I felt very dependent on my paper

I do my first CCE [customercourtesyescalation],the processfor impatientor

I take a call from an Army Base in their CareerResource Mgt. Center.The

PO [purchaseorder]numberthe caller wants to give me a very long "BPA"

During my two weeks on the phones, I remained dependent on my posted

knowledge,includingnotjust themechanicsof takingservicecalls, including

The "Quality Process": Quality or Control?

In contrastto the engineeringdesign world of negotiation,collaboration,

ideas vary,any good idea can always be reified andformula-izedas it moves

company.They also point out, to the company'scredit,thatit allowed QWL

same as practice.Process was formalizedand prescribedso thateven here it

The "Design Documents"

A series of documents illustrates the company's and the design program

|BUSINESS | | CUSTOMER CARE, PROBLEM RESOLUTION, TELEMARKETING, SUPPL

r SUPPLY ^ r TELE- ? r CUSTOMER '

-$ ORDERS I S N REVENUE CALLS FROM

Providefront.endproductsupport to customers through

*Service Requests *Pending Ser ice Calls

*HardwaceSertice Callneed Identified

*Service CaJlRecord *Service CallRecord

WORK PROCESS: SERVICECALL HANDLING9.3.1.1

Process Owner: Customer Services Support

Inputs* Service Requests

Key ProcessSteps: >^ \

\, reuiirmenti sSBQS\ Dispstch procesi

C ustomer , Field Technicians, Analysts

Service Call Record,Software Call Record,

this document, the representation of the detail is constrained by the inflexi-

and a visual representationof key process steps. A practice-orientedaction

XVORKPROCESS: SERVICECALL HANDLING9.3.1.1

Provide Front-end product support to customers through accuraet service

Process Owner: 8erviceCallHandling

Key Process Steps:

Harowarcservice call need Idenfified, Software call need identified,

Customer: Customer, Field Technicians, Analysts

nArnm-4o Service Call Record, 8offwarn Call Record,

Because of this change of communication format, some innovations based

!"#$%&'#$&($)+#,-'$./0#1+2$-3$+"#$4#2-53$6,&1#227$8$9&:;*,-2&3$&($!

Service Requests Pending Ser ice Calls

Service CaJlRecord Service CallRecord