1- hoschka : group "mobile" Mickael Johnston: map is a complex application, lots of info and so on. good application for multimodal to show what it is capable for other appli which are usually simpler, particularly classical business application. if possible for a map possible for lots of other appli. Ram - nokia: well the user does not really care for map appli, and thus not appealing MJ: well it is challenging appli, if you can make it for interaction with map you can do for lots of appli. Massimo: map are hard to make, and impressive, but we hsould look at problem and solution in // Ram: people should send their use case DSR: what about a questionnaire for application/use case : for which appli people would like to use multimodality 2- Mickael Johnston: group "other" Ram: more kinds of devices represented in cc/pp ? MJ: well easy to get phones description, not easy for other devices: printer/camera/... Stephane Boyera: even for phones, vendors are not releasing infos, major problem for cc/pp deployment and use General Discussion Debbie: interesting to see the mobile group didn't talk about diversity of mobiles Debbie: so discussin more on the short term ? Hoschka: more generic, not specific DSR: need/interest for short term simple multimodal authoring ML ? DSR: w3c could define some javscript to do script and bind that to xhtml interest ? hoschka: technology problem but also for what usage ? S sire: modulariztion of multimodal language ? DSR: well, we already doing that iwth eg inkml. so the problem is what time scale ? ram: now good opportunitu to define an auhtoring language instead of doing stuff post mortem like done for xhtml Debbie: mixed markup solution discusion in mobile ? Hoschka: not really DSR: so output of the WS: short term for auhtoring marukup. what is the scale : 4 years ? 2 years ? Ram: 4 years is long. well the fact is that now we have emma and inkml compare to 2002, that doesn't solve problem ??: problem of implementation ? Ram: well not take by the mass market. debbie: if there is a very important interst generally people take proprietary approach and the market proceed with it. maybe the proprietary solution didn't get deployed ? Ram: so do you think there is no user push but just technology ? DSR: my observation : lots of interest talking about MMI, but not real interest in application right now Debbie: what would be a multimodal authoring langage ? just mix voice with graphical ? or add other modalities ? short term not necessary ? J Sienel: what we need is synchronization between modalities. first approach graphical then speech. catching event on one modality make change on others. synchronisation is the key point massimo: smil designed to be an output langage, not input. lack of more general way to describe synchronization. just "second", not good for input. duration is important DSR: there is simple thing we can do in synchronized output.smil is good for that just need to look at details. S sire: no way to define multimodal event, like : i want to react on a "city name" enter MJ: composite input, which is handled by emma S sire: where to i put the event handler. how to i describe that i want ot react on that event and how i describe the event ? JS: definition of multimodal hanlder/event description ? DSR: all the point is to describe group of events. Vxml there is compounded event where you can specifiy just few parts DSR: just a w3c note at this point about that Debbie: need to proceed to mm event ? DSR: need to look further in that direction Debbie: dom3 events model. Debbie: the one thing on my list we didn't mention was haptics. Isn't anybody interested? Ram: still interesting for gaming DSR: yes, very simple uses ???: further discussion on the web ? dsr: a questionnaire will come max: yes public ML too debbie: summary ? hoschka: minutes to be published debbie: well high level conclusion are interesting. key output points