Graphical User Interface Complexity - Essay Example

Grphicl User Interfce (GUI) Complexity Introduction During the 1960s, we intercted with computers through punch crds nd pper, in the 1970s through online typewriters, in the 1980s through online video terminls, nd in the 1990s through multimedi terminls. In ll of these modes of interction, the user is primrily intended to be sitting t desk. In the 21st century, we re moving beyond the desktop nd intercting with computers in mny different environments nd through mny different technologies. In this pper, I will discuss those technologies nd the types of environments in which they will be useful. The discussion is bsed on two different perspectives. The explortion of the perspective of resercher of input/output devices who wishes to extend the design spce nd discuss n emerging prdigm of computer use, tsk-dependent computers, nd give some exmples of tht prdigm. Ultimtely, I will discuss the criticl role of input nd output devices, nd continue with the designer's perspective by discussing the motivtions for the evolution, the environments of use, nd the implictions of the environment on the interction technologies. The Criticl Role of Input nd Output Devices The mnner in which users interct with their computer is criticlly dependent on the choice of input nd output devices. This is such n obvious observtion tht sometimes system designers fil to tke it explicitly into ccount. ll interction with the computer consists of issuing commnd, providing dt on which the commnd will operte, nd, finlly, providing loction for the output of the commnd to be plced; this holds even when the execution of the commnd is temporlly disjoint from its invoction (e.g., scheduled opertion). The specifiction of those commnds involves specifying the commnd, the source of dt for the commnd, nd the trget for the output of the commnd. ll must be specified using vilble input nd output devices. Therefore, the choice of these devices will constrin the style of the interction possible for the user. (Bss et l., 2003:140) Consider implementing windows-style interfce, without pointing device such s mouse. The windows style depends on the user being ble to point to the source nd destintion of commnd. (Spitzer et l., 1997:49) Furthermore, the type of pointing device is importnt. The precision of pointing is not s importnt s the speed of pointing. Thus, device such s trckbll is possible pointing input device tht dmits to gret precision. Not only does the trckbll dmit gret precision, but lso demnds it. Thus, the use of trckbll is conditioned on the bility to very precisely specify loction on the screen, to serve s the current loction. Becuse precise specifiction tkes time, lower precision, but inherently fster pointing device, such s the mouse, will llow much smoother interction with the computer. (Bss, Mnn, Siewiorek, Thompson, 2002) The point of this discussion is tht there is coupling between the chrcteristics of n input or output device, nd the style nd mnner in which user cn operte computer. In this pper, we discuss environmentl nd tsk constrints on the choice of input nd output devices. It should be cler tht ny constrint on the devices my fundmentlly lter the type of interction tht the user hs with the computer. Input Devices In ny computer ppliction, distinction exists between input intended to control the computer (commnds) nd input intended to be retined (dt). The choice of pproprite input device will be hevily influenced by whether the device will be used primrily in the issunce of commnds or in the input of dt. In fct, the desktop hs seprte input devices for these two ctegories. The mouse is used primrily for commnds, wheres the keybord is used for the input of dt. The types of pplictions tht re primrily commnd oriented should not be discounted. ny dt retrievl ppliction (such s mny of those bsed on the World Wide Web) is primrily commnd oriented. (Normn, 2005) The commnds re used to nvigte to the desired dt item(s), nd then to specify how tht dt is to be displyed. In dt retrievl ppliction, only serch fcilities require free-text input. I will now briefly survey some input devices in light of the distinction between commnd nd free text input: Speech: In ny discussion of lterntive interction technologies, speech recognition is lwys prominent. Speech recognition is ppeling, in concept, becuse of the nturlness of speking. People re trined from birth to communicte using speech. Speech requires no hnds to operte. On the other hnd, speech s commnd-issuing mechnism is not prticulrly nturl. Dilogues re somewht stilted nd constrined by the vocbulry of the speech recognizer. The utility of speech s dt entry mechnism depends on the type of dt. If the dt re truly free text-grmmticl nd with rich context-then the error rtes of speech recognizers re sufficiently low nd dt cn be entered (nd, subsequently, retrieved) very efficiently. If the dt re of the "fill-in-the-form" (Normn, 2005) vriety-isolted words with no context-then speech recognizers generlly require n error correction mechnism nd Pointing devices: lthough the mouse is inherently desktop device, lrge number of lterntive pointing devices exist tht re suitble for two-dimensionl pointing. Joysticks or joypds cn be ttched to either the body or portble device, llowing rbitrry movement over two-dimensionl spce. If the device is ttched to the body, or on seprte fixed device, then it requires only one hnd to operte. If it is ttched on mobile device, then the number of hnds required depends on the size of the mobile device. For lrge device, one hnd is required to operte the pointer nd nother to hold the device stedy. For smll device, one hnd cn both hold the device nd operte the pointing device. Feedbck from pointing device is primrily visul-the cursor on disply-lthough uditory feedbck mechnisms cn be imgined. This type of device mkes the input of free text very wkwrd, but it is suitble for nvigtion or for the input of two-dimensionl dt, such s loctions on n imge (e.g., surfce or mp). pointing device must be used in conjunction with selection device, such s button, in order to specify tht the current loction is of interest. Pointing devices re useful for commnd-bsed input or for specifying loctions, but re poor for free-text input. (Normn, 2005) Selection devices: ny specil key cn be used s selection device. Buttons re the most common selection devices, lthough other devices, such s dil, cn lso be used for this purpose. The dil is used to nvigte round collection of known loctions on the output, nd the buttons surrounding the dil re used to select the current loction s being of interest. This device is used to nvigte through hierrchy, such s the World Wide Web, or through questionnire. Feedbck is usully visul, to indicte the current loction within the hierrchy. s with pointing devices, the input of free text is very wkwrd with selection device. Selection devices re useful for commnd-bsed interfces, but re poor when used to specify loctions, or for the input of free text. Pen: pen is device for the input of free text. It is used s pointing, selection, nd text input device. Text input requires the presence of text recognition softwre; the quite high qulity of recognition nd the low error rte of modern hndwriting recognition systems mke pen-bsed text input vible solution. Pens re typiclly used in conjunction with flt-pnel displys nd require two hnds to operte-one to hold the disply nd one to mnipulte the pen (lthough s we see lter, other types of pens re possible). Gesturing (hnds, hed, eyes): Gesturing includes input tht is generted through movements or position of the hed, the eyes, other portions of the body, or through objects being mnipulted by the user. ll uses of gesturing re for commnd input. For exmple, moving the hnd in prticulr mnner cn be mpped to prticulr commnd. Eye position cn be used to control the position of cursor. Gesturing input devices usully require some infrstructure in order for the gestures to be recognized. Gestures cn, in concept, be used s pen replcement, but, in prctice, they re used for commnd input nd not for free-text input. The choice of prticulr input device depends on the environment, the tsk to be ccomplished, nd the cpbilities of the user. It is lso necessrily coupled to the choice of output device. s we indicted, feedbck from input is usully required, nd this feedbck is presented through the output device. We return to this point when we discuss the construction of system from input nd output devices. Output Devices We discuss output devices bsed on the modlity/humn sense they pply to. Three of the senses re commonly considered for recognition of computer output: sight, sound, nd touch. Of these, the most commonly used for computer output is sight. Detils of these types of output devices follow: Sight: The visul devices designed for non-desktop computer usge re flt-pnel displys nd hed-mounted displys. Flt-pnel displys require either hnd to hold them, or convenient plce within the viewing spce to mount them. lthough, in concept, mgnifier could be plced in front of flt-pnel disply nd the physicl size of the disply mde quite smll, in prctice these displys usully present n imge where the pprent nd the rel sizes re identicl. Hed-mounted displys, on the other hnd, rely on the distinction between rel nd pprent sizes. very smll disply (liquid crystl in current technology) is reflected through the use of optics nd mde to pper much lrger. Hed-mounted displys cn be monoculr or binoculr. Binoculr displys re used for ugmented relity pplictions where computer imge is imposed over the imge the user would ctully see without the disply. Monoculr displys re used s informtion displys, without ny reference to the view the user would ctully see. They re used, in essence, s mobile desktop displys. Becuse of the mnner in which they re positioned, the user cn look round them, or t them, much s bifocl glsses re used. Hed-mounted displys suffer from problem of socil cceptnce. Current displys re lrge nd cumbersome, nd give the werer the ppernce of figure from science fiction movie. These devices re becoming smller nd less obtrusive, however. udio: Sound, both spoken nd non-spoken, is lso used s n output device. Speech cn be used to replce textul disply, s well s for feedbck for vrious forms of input. The use of sound to replce textul disply involves not only the presenttion of the words, but lso djustment of the presented informtion becuse of differences between the temporl properties of sound versus the sptil properties of disply. Text shown on disply will remin there until it is removed, either by the system or by the user. Hence, the speed of understnding is t the control of the user. udio output hs short durtion nd must be understood when it is spoken, or it must be repeted. Hence, the speed of understnding is t the control of the system. This leds to differences between crrying out dilogues with systems using visul output nd with systems using udio output. udio output is lso sensitive to mbient noise. Non-speech sound hs been used both for lrms nd for directionl indictions. Its use for direction hs minly ddressed the needs of the blind. Using combintion of route known by the computer nd positionl loction system, such s the Globl Positioning System, sound is used to inform users tht they hve devited from the plnned route. By sending the sound to the right or left er, s pproprite, users cn be informed not only of the fct tht they hve devited from the plnned route, but lso of the direction in which they should turn. Tctile: The use of tctile output for directionl indictors hs been suggested for drivers of utomobiles. Tht is, the set or steering wheel will inform the driver of the direction of the next turn. This suggestion hs, s yet, not been implemented. The sme is true of existing tctile technology tht is currently used to present the output of computing systems to blind users in n ttinble form, but hs not yet found ppliction in more minstrem systems. System Design It is in the design of the system tht the considertions we hve discussed so fr come together. The style of the possible interctions is dependent on the choice of input nd output devices; the choice of devices is, in turn, dependent on the constrints imposed by the environment nd the tsk of the user. (Ksbch et l., 2004:175) Of the devices tht we hve discussed, the only inherent coupling between input nd output devices is between the pen, s we hve described it, nd the flt-pnel disply. In lter section, we see tht even this coupling is not necessry. In ny cse, using current-or soon to be vilble-technology, the system designer mkes choices bsed on the tsk, the number of hnds necessry to operte the input devices, the senses vilble from the user when operting the input device, the error rte of the device, nd the extent to which trining is required either to tilor the device to the user or to trin the user to the device. The growth of tsk-dependent computers is bsed on the underlying technology becoming "good enough" (Ksbch et l., 2004:174) so tht specil-purpose devices cn be designed nd constructed. The interction technologies ssumed by these devices re tightly coupled input nd output devices, but limited in both spects. Thus, the cler trend is to specilize, to scrifice generlity nd to utilize the technology in the service of the user, rther thn force the user to dpt to few limited forms of interction. In ddition, the tsk-, user-, nd context-wre devices of the future will introduce new cpbilities in ddressing the specific needs of the end-user popultion, which constntly increses in size nd diversity. To tke dvntge of these opportunities, designers will hve to cquire better understnding of the design spce of input nd output devices, s well s of the implictions tht specific choices/combintions in tht spce hve on interction. Furthermore, by directing reserch efforts to the development of "trgeted" interfces for specific tsks nd contexts of use, it is possible to constrin the intrinsic complexity in the development of user interfces tht re ccessible nd usble by ll. Bibliogrhy: 1. Bss, L., Ksbch, C., Mrtin, R., Siewiorek, D., Smilgic, ., & Stivoric, J. (1997). The design of werble computer. In Proceedings of the Humn Fctors in Computing Systems Conference (pp. 139-146). New York: CM Press. 2. Bss, L., Mnn, S., Siewiorek, D., & Thompson, C. (1997). Issues in werble computing: CHI '97 workshop. SIGCHI Bulletin, 29 (4). [Online]. vilble: http://www.cm.org/sigchi/bulletin/1997.4/bss.html 3. Fishkin, K. P., Morn, T. P, & Hrrison, B. L. (1999). Embodied user interfces: Towrds invisible user interfces. In P. Dewn & S. Chtty (Eds. ), Proceedings of the IFIP Working Conference on Engineering for Humn-Computer Interction (pp. 55-73). Norwlk, CT: Kluwer. 4. Hrrison, B. L., Fishkin, K. P., Gujr, ., Mochon, C., & Wnt, R. (1998). Squeeze me, hold me, tile me! n explortion of mnipultive user interfces. In Proceedings of the Humn Fctors in Computing Systems Conference (pp. 17-24). New York: CM Press. 5. Ksbch, C., Pcione, C., Stivoric, J., Gemperle, F., & Siewiorek, D. (1998). Digitl ink: fmilir ide with technologicl might! In Proceedings of the Humn Fctors in Computing Systems Conference (pp. 175-176). New York: CM Press. 6. Normn, D. (1998). The invisible computer. Cmbridge, M: MIT Press. 7. Spitzer, M. B., Rensing, N. M., McClellnd, R., & quilino, P. (1997). Eyeglss-mounted displys for werble computing. In L. Bss & . Pentlnd (Eds. ), Proceedings of the 1st Interntionl Symposium on Werble Computers (pp. 48-51). Cmbridge, M: IEEE Press. vilble: http://www.computer.org/conferen/proceed/8192/8192toc.htm. Read More