CSCW Issues and Theory Ov erview computer systems ha v e - PDF document

CSCW Issues and Theory Ov erview computer systems ha v e group impact A l l not just group w are : : : Ignoring this leads to the failure of systems W e lo ok at sev eral lev els, from min utiae to large scale con text: face-to-face comm unication � con v ersation � text based comm unication � group w orking � organizational issues � Human{Com puter In teraction, Pren tice Hall Computer{Supp orted Co op erativ e W ork (1) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 14

F ace-to-face comm unication Most primitiv e and m ust subtle form of � comm unication Often seen as the paradigm for � computer mediated comm unication? T ransfer e�ects carry exp ectations in to electronic media � sometimes with disastrous results ma y in terpret failure as rudeness of colleague e.g., p ersonal sp ac e video ma y destro y m utual impression of distance happily the `glass w all' e�ect helps Human{Com puter In teraction, Pren tice Hall Computer{Supp orted Co op erativ e W ork (2) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 14

Ey e con tact to con v ey in terest and establish so cial presence � video ma y sp oil direct ey e con tact � (recall video tunnel, Ch. 13) but p o or qualit y video b etter than audio only � Gestures and b o dy language m uc h of our comm unication is through our � b o dies gesture (and ey e gaze) used for deictic r efer enc e � head and shoulders video loses this � close fo cus for ey e con tact So : : : or wide fo cus for b o dy language? Human{Com puter In teraction, Pren tice Hall Computer{Supp orted Co op erativ e W ork (3) � c A. Dix, J. Finla y , G. Ab o wd and R. Beale 1993 Chapter 14

Bac k c hannels 1 Do y ou fancy that �lm Alison: err : : : : : : 2 `The Green' um : : : it starts at eigh t. Great! Brian: Not just the w ords! Bac k c hannel resp onses from Brian at 1 and 2 quizzical at 1 a�rmativ e at 2 Bac k c hannels include: no ds and grimaces shrugs of the shoulders grun ts and raised ey ebro ws Utterance b egins v ague then sharp ens up enough just Human{Com puter In teraction, Pren tice Hall Computer{Supp orted Co op erativ e W ork (4) � c A. Dix, J. Finla y , G. Ab o wd and R. Beale 1993 Chapter 14

Bac k c hannels I I Restricti ng media restricts bac k c hannels video | loss of b o dy language audio | loss of facial expression half duplex | lose most v oice bac k c hannel resp onses text based | nothing left! Bac k c hannels used for turn-taking: sp eak er o�ers the �o or � (fraction of a second gap) listener the �o or r e quests � (facial expression, small noise) Grun ts, ` um 's and ` ah 's, can b e used b y the: listener to the �o or claim � sp eak er to the �o or hold � but often to o quiet for half-duplex c hannels T rans-con tinen tal conferences { sp ecial problems lag can exceed the turn taking gap � leads to a monologue! Human{Com puter In teraction, Pren tice Hall Computer{Supp orted Co op erativ e W ork (5) � c A. Dix, J. Finla y , G. Ab o wd and R. Beale 1993 Chapter 14

Basic con v ersational structure Do y ou fancy that �lm Alison: the ( 500 ms ) with the blac k cat Brian: uh | `The Green whatsit' y eah, go at Alison: uh : : : ( lo s ) : 20 to? oks at watch | 1.2 : : sure Brian: Smallest unit is the utter anc e T urn taking utterances usually alternate = ) Simplest structure | adjac ency p air Adjacency pairs ma y nest; Do y ou w an t some gateau? Brian: is it v ery fattening? Alison: y es, v ery Brian: and lots of c ho colate? Alison: masses Brian: I'll ha v e a big slice then. Alison: Structure is: A-x, B-y , A-y , B-z, A-z, B-x Inner pairs often for clari�cation But, try analysing the �rst transcript in detail! Human{Com puter In teraction, Pren tice Hall Computer{Supp orted Co op erativ e W ork (6) � c A. Dix, J. Finla y , G. Ab o wd and R. Beale 1993 Chapter 14

Con text in con v ersation Utterances are highly am biguous W e use to c ontext disambiguate ( p oints ) that p ost is leaning a bit Brian: that's the one y ou put in Alison: Tw o t yp es of con text: external con text reference to the en vironmen t e.g., Brian's ` that ' | the thing p oin ted to deictic r efer enc e in ternal con text reference to the previous con v ersation e.g., Alison's ` that ' | the last thing sp ok en of Often con textual utterances in v olv e als : indexic that, this, he, she, it these ma y b e used for in ternal or external con text Also descriptiv e phrases ma y b e used: external: `the c orner p ost is le aning a bit' in ternal: `the p ost you mentione d' Human{Com puter In teraction, Pren tice Hall Computer{Supp orted Co op erativ e W ork (7) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 14

Common Ground Resolving con text dep ends on meaning � = participan ts m ust share meaning ) so m ust ha v e shared kno wledge Con v ersation constan tly negotiates meaning � pro cess called gr ounding So, y ou turn righ t b eside the riv er. Alison: past the pub. Brian: y eah Alison: : : : Eac h utterance is assumed to b e: � | furthers the curren t topic r elevant | comprehensible to listener helpful Human{Com puter In teraction, Pren tice Hall Computer{Supp orted Co op erativ e W ork (8) � c A. Dix, J. Finla y , G. Ab o wd and R. Beale 1993 Chapter 14

F o cus and breakdo wn Con text resolv ed relativ e to curren t dialo gue fo cus Oh, lo ok at y our roses Alison: : : : mmm, but I'v e had trouble with green�y . Brian: they're the sym b ol of the English summer. Alison: green�y? Brian: no roses silly! Alison: T racing topics is one w a y to analyse con v ersation. Alison b egins | is roses topic Brian shifts topic to green�y Alison misses shift in fo cus br e akdown : : : Breakdo wn happ ens at all lev els: topic, indexicals, gesture Breakdo wns are frequen t, but r e dundancy mak es detection easy (Brian cannot in terpret ` they'r summer ') e : : : p eople v ery go o d at r ep air (Brain and Alison quic kly restore shared fo cus) Electronic media ma y lose some redundancy = breakdo wn more sev ere ) Human{Com puter In teraction, Pren tice Hall Computer{Supp orted Co op erativ e W ork (9) � c A. Dix, J. Finla y , G. Ab o wd and R. Beale 1993 Chapter 14

Sp eec h act theory A sp eci�c form of c onversational analysis � Utterances c haracterised b y what they do , � they are acts : : : e.g., `I'm h ungry' | h unger pr op ositional me aning in tended e�ect | `get me some fo o d' Basic con v ersational act the il lo cutionary p oint : promises, requests, declarations, : : : Sp eec h acts need not b e sp ok en e.g., silence often in terpreted as acceptance : : : Generic patterns of acts can b e iden ti�ed Con v ersation for action (CfA) regarded as cen tral Basis for group w are to ol Co or dinator structured email system � users m ust �t within CfA structure � not lik ed b y users! � Human{Com puter In teraction, Pren tice Hall Computer{Supp orted Co op erativ e W ork (10) � c A. Dix, J. Finla y , G. Ab o wd and R. Beale 1993 Chapter 14

Con v ersations for action A:Decline A:Request B:Promise B:Assert A:Declare 1 2 3 4 5 A:Counter A:Accept B:Renege B:Counter B:Reject 6 7 A:Withdraw A:Withdraw A:Withdraw A:Reject B:Withdraw 8 9 Circles represen t `states' in the con v ersation � Arcs represen t utterances (sp eec h acts) � Simplest route 1{5: Alison: ha v e y ou got the mark et surv ey on c ho colate mousse? request Brian: sure promise Brian: there y ou are assert Alison: thanks declare More complex routes p ossible, e.g., 1{2{6{3 : : : Alison: ha v e y ou got request : : : Brian: I'v e only got the summary �gures coun ter Alison: that'll do accept Human{Com puter In teraction, Pren tice Hall Computer{Supp orted Co op erativ e W ork (11) � c A. Dix, J. Finla y , G. Ab o wd and R. Beale 1993 Chapter 14