Minutes MMI workshop, 19 July, afternoon Philipp Hoschka "Multimodal Applications from Mobile to Kiosk" ---------------------------------------------- Michael Johnston, ATT Kirusa: what is intrinsic property of comp inpt whch prevents you from splitting them into two inputs ? take example of cirling A: depends on complexity of application - single piece of ink/spoken comannd may be true. but not ifyou have multiple pieces of ink and multiple pieces of speeck "Vodafone Position Paper" ------------------------- Maria Farrugia, Vodafone Q: how do measure easiness ? easy compared to what ? A: we run focus groups, get feedbac from our services - weatherforecast - number of keypresses - we know where the barriers are - look at tech and use them to overcome barrieres - use user groups to measure - quantative analysis afterweards - hits to wehater now compared to after (massimo) Q: most servie are about location based - user position is sensitive information - are users aware of that ? A: in the process of investigating this - tryig to get feeeback - what user perceives as sensitive information - whether people would use it Fabio: what kind of environment have ou used - what implementation environment ? A: not involved in implementation myself Fabio; publically avaialbe ? A: not available yet - i am in R_D Alcatael: what about new services ? automatically propagated ? new servide in vodafone life - how to propagate ? complete client implemetatin A: demo is client implementation0 there are way of doing int in vodafone life - when page comes up text that covers links automatically speech enabled - recognition done inn handset Q: (tuolouse) - screen size diversity - what strategy do you use to cope with this ? A: big problem for us - want ot target business customers with big screens and others - stratgy is same way as OMA - device capabliity sent to interaction manager which automatically adopts content to device - if screen not big enough, render info via speech "SAP position paper" -------------------- Cedric Ulmer, Anne Hardy, SAP Q: RML specification - ... -- how does it work ? A: reverse engineering not the approach- authors have to develop in RIML - basing it on XHTM, XForms and SMIL would ease it for developers was assumption Q: fabio - what rules for transformation ? A: there is automatic part - layout adaptation - pagination - you can influence it - task pagination we dont do - pagination based on layout - no task modelling - we do content selection Kirusa: security applications ? .. A: firefigthers use a lot of tme - may come too late - fires grow very quicly in the area hiere = for now, just walkie-talkies - -based on ERS, they have trucks, and teams - need to know where are my trucks, are they in the area - is my team on holiday - that's the ERPC side Q: in RIML, what do you use SMIL for ? A: content selection, not for synchronisation "Component Model for MMI" -------------------------- Dave Raggett, W3C/Canon Q: (Toulouse guy) bidning mechanism - do you think candidate for recommendation of itself, or emebedeed ? Looks as if this is embedded in SVG as RCC A: interest in several groups- SVG is a bit ahead, so they may end up owning it - but more generally useable Q (?): decouple ... A: different UIs for different devices and different user preferences Q: if application logic is tied to devices - logic may be different from different devices A: need OO design skills - figure out UI - how to represent them on devices - chunkingissue - maybe you can hide thes eissues in UI Q: (ETSI) user profiles - can this platform cooperate with user profile ? not only find info about my screen, but also diffretn things - say i'm colorblind A: desccription of device capalibilies et. is what S+E is about - selecting speech only application e.g. A (StephaneB): we are developing framework for context: device characteristics, user preferences and environment - this is what CC/PP solves - "structures and vocab" is now a W3C recommendation - focussing on protocol and processing rules part now, in cooperation with S+E "USIXML: a User Interface Description Language for Specifying Multimodal User Interface" ---------------------------------------------------------------------------------------- Jean Vanderdonckt, Universite Catholique de Louvain (presented by Daniela Trevisan) Q: (??) ery flexible toolset - how do you port from HTML to another format ? do you have to write translation rules ? A: if you use tools, will produce usixml code - interchange uses usixml definition - can start from HTML - tool will convert into USIXML - assume you want to do a new interface for mobile phone - an resue transformation rules - or develop your own using the graph grammar "The Markup Way to Multimodal Toolkits" --------------------------------------- Stephane Sire, Intuilab Q: (Fabio) need for more flexible control flow - what do you mean ? A: model is not client server with one control flow - common model in browser based interaction - in my model, event can be dispatched to different parts that have their own state management - several threads running in parallel - calling for more control structures that allow local event loops - that's what we do with finite state machines - events can be received asynchronously Q: (Dave) is this concurrent state machines ? A: yes - havent' investigatge to share same statem chancine - would maybe use for petri nets - seems more appropriate - we have stopped just before the petri net stage - in the architecture i describe, there's a core model - comebody can add a petri net model - "Modularization of Multimodal Interaction Specifications" --------------------------------------------------------- Matthias Denecke, NTT (no questions)