Designing user interfaces—the role in intuition and imagination (1992)

It is argued that too little Is known about the cogn]tl} e aspccls of dcslgn. This knowledge IS essent]al if lhe man! guldclincs, models and tools tha{ ha\e emerged ]n the f]eld of user Inwrfacc dcslgm arc to ha} c a signiflcan[ impact on dcslgn pract]cc. Emplnctd studies of dcslgncrs de~cloplng user lntcrfaccs arc reported, showng that the context in whtch design takes place !n an organlsattonid semng IS (urbulen[ and the des]gn task (Jl[cn unclear. ln\ estlgal]mrs ol dects]on makng in the design pr(xess sh(mcd II IS one of graduallj cioll Ing comm[tmcnt. w here Intulllon, lmwynauon and unswucturcd wral!sls urc csscnuul cognllllc processes during ctcslgn u (wk. The u~cfulncs~ ol dcslgncr supfxm tmi~ IS dlscusscd. Ittlrodm Iitw. C(~gn]ti\ c aspcc[~ of Ihc dcslgn procms ha\c been ln~ cwgatcd in a number of studlcs of dccls]on making in user In[crfacc dcslgn (c, g, BcltMII 1990, Jorgensen 1990). The studies were rctr(lspcctlt c and based on Inten iews or questionnmrcs The stud]cs ha$ e been \aluable to the HC1 communtt> how cl cr, the! do not tell us w hat design IS all alxmt and h(~w des]gncrs reason during the actual dcslgn pr~wcss II has been argttcd tha[ design problems arc ambiguous (hfdh[)tra 1%30) and III dcl’lncd, and u III thcrclorc rcslst formal anal! lical rncthods (Klein 1987). The speclflc ct)gnllltc quullflcit[t{)ns of (hc cics]gners wc of kc) ]mpfvlanc’c. H(n\c\ cr, the dcl cl(~pmcn[ of [hconcs of dcwgncrs’ ct>gnllll c pr(wcsw$ arc prcrcqulsltcs Rcpmmng on LI number [~f d] ffcrcn[ ln~cstlgutlt>n~ t>f s(~lut}{}n gcncrati(m ti> gl) cn dcslpn problems Turner ( 1987) concludes that “.. the cognltl\ c prcxcsscs }n~ 011cd In dcs]gn seem to be iLswx-IaIIJ c and Indlt Idual... ”, and he argues that lIItlc IS knou n about u hii( destgn IS or how people go abou~ d(>lng II earned out through the AMODEUS project, dcm.rmentln~ the organisational context of a design process. A designer w’as obsemed for 16 \vork days over a period of 7 months during the de! ’elopment of a database system. Task anal) sis. screen design, protot}pe demonstrations and meetings if ith users were obsen ed. The ltwestigatlons showed lhat the organisational ccm(exi \vas highly turbulent, the u ork setting was characterized by frequent interruptions, and the task mformatiorr al ailable was fragmented, uncertain, and incomplete. The user Interface strategy becarnc ‘muddl}ng through’ (Jorgensen et al. 1991), an opportumst!c design strategy described by Gtundon ( 1990), lvh~ch provides evidence on the complci and polymorphic character of dccis]on making ]n design, These obsen ations point at the influence of the organisat]onal conte~t on the design process and it should be taken into account \vhcn trying m conceptualize the design process. With the aim of gaimng lns]ght Into deslgncrs’ dcclslcm makng, e~penmental workshops wclth experienced designers working Mlth user Inter-face design were also earned out, The designers worked wth a design task for approximate> 1 hour u hl Ie bmng I deotapcd. The desl gn problem was h(~il LOdes]gn a user interface for a \ ]deo picture database The process anid! SISshe\\s an ]tcratl\e work process, \vhcrc ~ctf e\pl]c]t decls]ons are made. the process IS rather onc of gradually e}olving commitment, nhcrc intuition, imagination and unstructured analysis based III clpcncncc arc the csscntlal cogn]t]~ c quallflwtioni at cfcwgn N ork, Thcj seem to bc the dri\ ing force in the prooxs (Nlclscn 1991). Rational analytical thlnklng enters onl> secondl~ and seems to function more as a tool for checking out (he lmag]ned design, \erilying or falslfytng It during simulated runs (Adelson et al 1988). T& usogtI ojwpi)ort [OOISitl the de~ig)l proces ~ Many different systems and user models ha~e been de~ckyxxi to support and structure the des]gn process In order to make gains m softl~ are producmt it> and qual]tj. HCM{c1er, the usage c)f such models b! dcslgners has not been s() successful w e\~cted, The models, u hich ]ns]s[ on a more or less rigid structure from the outset, secm too complc~, m)t I CD usable in real w or-k set[ings, and designers often get lost In the detiul w hcn tr-j ing to appl> these techmqucs. As an example, the Design Rationale approach IS suggested to be used as a way of analywng the content of a dcslgn session. The Design Rationale Frameu ork cons! ~lln~ of three ma]n concepts. Questions, Opt]ons. and Crltcr]ti. can be seen as a techmque for ca~lng out des]gn, and helps w understand why the design is the w a} It IS, to understand the problcm better and therefore produce mot-c e( IcctI\ c sol utions (hLicLean 1992),