IRC log of multimodal on 2016-06-06

Timestamps are in UTC.

11:50:06 [RRSAgent]
RRSAgent has joined #multimodal
11:50:06 [RRSAgent]
logging to
11:50:14 [kaz]
rrsagent, bye
11:50:22 [Zakim]
Zakim has joined #multimodal
12:58:04 [ddahl]
ddahl has joined #multimodal
13:06:49 [kaz]
Meeting: MMI
13:06:54 [kaz]
scribenick: kaz
13:07:15 [kaz]
Chair: Debbie
13:07:24 [kaz]
present: Debbie, Helena, Dirk, Kaz
13:09:17 [kaz]
topic: Actions
13:09:20 [kaz]
13:09:20 [trackbot]
action-440 -- Kazuyuki Ashimura to Tell wot ig about our bullet points for the technology landscape -- due 2016-05-23 -- OPEN
13:09:20 [trackbot]
13:09:43 [kaz]
kaz: talked with the WoT IG
13:09:50 [kaz]
... and got OK to submit the bullet points on MMI
13:10:07 [kaz]
... still need to submit the actual text to GitHub using pullrequest
13:10:35 [kaz]
close action-440
13:10:35 [trackbot]
Closed action-440.
13:11:03 [kaz]
action: Kaz to submit bullet points on MMI to the WoT Landscape document on GitHub
13:11:03 [trackbot]
Created ACTION-441 - Submit bullet points on mmi to the wot landscape document on github [on Kazuyuki Ashimura - due 2016-06-13].
13:11:40 [kaz]
ddahl: explains the above action to Dirk
13:11:53 [kaz]
... liaison with WoT IG
13:11:58 [kaz]
... Kaz will submit the text
13:12:10 [kaz]
topic: Use Case Wiki
13:14:16 [kaz]
-> Use Case wiki
13:14:24 [helena]
helena has joined #multimodal
13:14:41 [dirk]
dirk has joined #multimodal
13:14:46 [kaz]
ddahl: Helena and Dirk met before?
13:14:48 [kaz]
helena: no
13:15:12 [kaz]
... Helena working on MMI discovery
13:15:34 [kaz]
... now CTO of an enterprise company
13:15:58 [kaz]
dirk: post-doc research on mmi an ubiquitous
13:16:26 [kaz]
... also developed opensource voice system
13:16:49 [kaz]
... trying to support more modalities for car infortainments
13:16:58 [kaz]
... focused on speech
13:17:23 [kaz]
ddahl: the other topic I wanted to talk was a CG on voice
13:17:34 [kaz]
... use case and requirements for new voice applications
13:17:43 [kaz]
... like virtual assistance
13:17:55 [kaz]
... there should be newer standards
13:18:25 [kaz]
... CGs give us opportunity to develop specifications thought they're not official W3C recommendations
13:19:18 [kaz]
... put into a mission statement to launch the CG
13:19:37 [kaz]
... not heavy IP for CG
13:20:39 [kaz]
kaz: add some note on the IP things on CGs
13:21:09 [kaz]
ddahl: would like to try new things like SLM
13:21:34 [kaz]
... also probably some ideas on dialog management, e.g., user-initiated ones
13:21:48 [kaz]
dirk: state-based dialog management, etc.
13:22:03 [kaz]
... how that can be standardized
13:22:20 [kaz]
ddahl: ok
13:22:33 [kaz]
... let's go back to the Use Case discussion
13:22:38 [kaz]
topic: Use Case Wiki
13:22:44 [kaz]
-> Use Case wiki
13:23:02 [kaz]
ddahl: would like to see Dirks comments
13:23:09 [kaz]
... esp. the first two comments
13:23:23 [kaz]
-> comment 1
13:24:23 [kaz]
ddahl: regarding UC-7
13:24:47 [kaz]
-> UC-7
13:25:27 [kaz]
-> UC-7
13:25:35 [kaz]
s|-> UC-7||
13:26:04 [kaz]
dirk: multimodal approach to control your home
13:26:11 [kaz]
... we had SCXML runtime
13:26:22 [kaz]
... interacting different standards for presentation
13:26:30 [kaz]
... VoiceXML for speech
13:26:41 [kaz]
... VML for avator
13:26:55 [kaz]
... EMMA as data
13:27:14 [kaz]
13:27:33 [kaz]
s/BML/BML (Behavior Markup Language)/
13:27:51 [kaz]
... user can control device within home
13:27:57 [kaz]
... this project has ended
13:28:24 [kaz]
-> Dirk's paper
13:28:41 [kaz]
dirk: we use a lot of standards
13:28:52 [kaz]
... only few developers who know all of them
13:29:18 [kaz]
... how to discover devices and their capabilities
13:29:31 [kaz]
... dynamically
13:29:54 [kaz]
... coming from academic approach
13:30:02 [kaz]
13:30:07 [kaz]
... we took some shortcuts
13:30:36 [kaz]
ddahl: gap section
13:31:16 [kaz]
... trying to sign-in
13:31:44 [kaz]
... gap or general problems for developers
13:32:46 [kaz]
(updating the wiki)
13:33:16 [kaz]
ddahl: probably the discovery draft would handle the first question
13:34:09 [kaz]
(updated the Gaps section of the UC-7)
13:35:11 [kaz]
dirk: need more data
13:35:36 [kaz]
... the user say "open the shutter"
13:35:52 [kaz]
... and the system need to identify which shutter should be opened
13:36:06 [kaz]
... upper Interaction Manager for MMI fusion
13:36:30 [kaz]
... serious problem with tree structure
13:37:16 [kaz]
ddahl: we never really clear about what the fusion engine would fit
13:37:36 [kaz]
... some people mentioned we needed a separate fusion engine
13:37:47 [kaz]
dirk: that's a possible opton
13:37:51 [kaz]
13:38:13 [kaz]
... if we have more complicated system, that would be tricky
13:38:49 [kaz]
kaz: Prof. Araki and his Team had some idea about how to handle fusion/fission
13:39:07 [kaz]
ddahl: yeah
13:39:33 [kaz]
... also need for additional API or events for fusion/fission?
13:39:55 [kaz]
dirk: we rely on a rule engine
13:40:11 [kaz]
... therefore EMMA within data object was good
13:40:36 [kaz]
... but maybe some dedicated fusion/fission tags could be added
13:40:45 [kaz]
... we should look into fusion/fission capability
13:40:52 [kaz]
13:40:55 [kaz]
13:41:12 [kaz]
ddahl: probably we need something more than traditional lifecycle events
13:41:45 [kaz]
kaz: kind like "merge cells" and "split cells" for MS Excel :)
13:42:02 [kaz]
dirk: fusion/fission needs context information
13:42:17 [kaz]
... need some easier access mechanism
13:43:10 [kaz]
... can handle it using extension event, though
13:43:58 [kaz]
kaz: wanted to ask about what kind of extension was added to EMMA and MMI lifecycle events
13:44:20 [kaz]
dirk: introduced additional component to the tree structure
13:44:42 [kaz]
kaz: did you extend EMMA?
13:44:52 [kaz]
dirk: no, EMMA was sufficient
13:45:05 [kaz]
... just introduced compound XML forat
13:45:10 [kaz]
13:46:30 [kaz]
kaz: did you use additional events in addition to the usual MMI lifecycle events?
13:47:00 [kaz]
dirk: no, data was bypassed
13:47:27 [kaz]
... share the context
13:47:40 [kaz]
s/share/to share/
13:47:56 [kaz]
... extensionNotification was not very useful for our system
13:48:16 [kaz]
kaz: ok
13:48:32 [kaz]
... so we should think about that "bypass" mechanism too
13:48:36 [kaz]
dirk: bypassing event chains
13:48:56 [kaz]
ddahl: something like a "blackboard"?
13:49:01 [kaz]
dirk: yes, kind of
13:49:23 [kaz]
... context information could be shared using that mechanism
13:49:44 [kaz]
ddahl: MCs can access the blackboard?
13:49:54 [kaz]
dirk: no, only the fusion/fission manager can
13:50:18 [kaz]
ddahl: like the terminology of "upper Interaction Manager"
13:51:01 [kaz]
helena: control layer?
13:51:18 [kaz]
... we need a high-level control layer
13:51:55 [kaz]
... IM can't handle every aspect
13:53:50 [kaz]
... the data component should provide contextual information
13:54:05 [kaz]
dirk: data component within the MMI Architecture?
13:54:13 [kaz]
ddahl: part of the IM?
13:54:51 [kaz]
dirk: share the data to made it accessible by the upper IM
13:55:33 [kaz]
helena: as a part of the control layer, you can store the data
13:55:45 [kaz]
... you can put into the nested structure
13:57:05 [kaz]
-> Runtime architecture diagram
13:57:10 [kaz]
ddahl: we're running out of time
13:57:18 [kaz]
... great if Dirk can come back again
13:58:01 [kaz]
kaz: yes
13:58:06 [kaz]
dirk: ok
13:58:26 [kaz]
... btw, there will be another SCXML Workshop in Brussels in 2 weeks
13:58:36 [kaz]
... June 21
13:58:47 [kaz]
... one-day workshop
13:59:27 [kaz]
-> SCXML Workshop site
14:00:00 [kaz]
rrsagent, make log public
14:00:06 [kaz]
rrsagent, draft minutes
14:00:06 [RRSAgent]
I have made the request to generate kaz
14:04:22 [kaz]
i/Actions/This week Dr. Dirk Schnelle-Walka from Harman International, Connected Car Division joined us to discuss MMI use cases./
14:04:25 [kaz]
rrsagent, draft minutes
14:04:25 [RRSAgent]
I have made the request to generate kaz
14:14:04 [ddahl]
s/developed opensource voice system/developed open source VoiceXML system
14:15:04 [ddahl]
s/we never really clear/we were never really clear
14:15:30 [ddahl]
s/also need for/also was there a need for
14:16:32 [ddahl]
s/like the terminology of "upper Interaction Manager"/I like the terminology of "upper Interaction Manager"
14:17:00 [ddahl]
rrsagent, draft minutes
14:17:00 [RRSAgent]
I have made the request to generate ddahl
14:17:41 [ddahl]
rrsagent, format minutes
14:17:41 [RRSAgent]
I have made the request to generate ddahl
14:26:58 [ddahl]
s/the user say "open/the user says "open
14:27:22 [ddahl]
s/what the fusion engine would fit/where the fusion engine would fit
14:28:01 [ddahl]
rrsagent, format minutes
14:28:01 [RRSAgent]
I have made the request to generate ddahl
14:54:26 [ddahl]
ddahl has left #multimodal
16:00:10 [Zakim]
Zakim has left #multimodal
16:29:27 [kaz]
kaz has joined #multimodal