See also: IRC log
This meeting concludes our initial round through mandatory requirement spec proposals. The basic idea for a modality tag is clear; Enrico will ask specific questions to the group to assess what people may need.
Paolo mentioned an alternative solution for defining the set of attribute values which we need in many places (Core 1, Core 2, etc.) The basic idea is to use QNames in the attribute value; he will circulate a link to another language where this is used.
For Confidence, there seems to be agreement to use is on core emotion annotations (Core 2-5, Core 7); more thinking is needed to decide whether it should also be used for complex emotions (Core 6) and global metadata (Global 0).
The meeting in Cannes will walk through all topics, but focus on improving agreement on Links and Global Metadata. The people present at the f2f will work out concrete proposals to be discussed further via the mailing list.
<marc> Enrico's proposal: http://lists.w3.org/Archives/Public/public-xg-emotion/2008Aug/0016.html
Enrico introduces his proposal wrt. Meta 2 - Modality
Enrico explains sign as stemming from sign languages
<marc> <emotion>
<marc> <category set="everyday" name="pleasure"/>
<marc> <modality medium="visual" set="basic_modalities" mode="face">
<marc> </emotion>
<marc> <emotion>
<marc> <category set="everyday" name="pleasure"/>
<marc> <modality medium="visual">
<marc> </emotion>
<marc> Multimodality, suggestion 1:
<marc> <emotion>
<marc> <category set="everyday" name="excited"/>
<marc> <modality id="m1" set="basic_modalities" mode="face"/>
<marc> <modality id="m2" set="basic_modalities" mode="voice"/>
<marc> </emotion>
<marc> suggestion 2:
<marc> <emotion>
<marc> <category set="everyday" name="excited"/>
<marc> <modality set="basic_modalities" name="face&voice"/>
<marc> </emotion>
<marc> suggestion 3:
<marc> <emotion>
<marc> <category set="everyday" name="excited"/>
<marc> <modality set="basic_modalities&my_modalities" name="voice&music"/>
<marc> </emotion>
Enrico names further example by Felix
<marc> suggestion 4:
<marc> <modality set="basic_modalities">
<marc> <mode="face"/>
<marc> <mode="voice"/>
<marc> </modality>
<paolo> Felix email: http://lists.w3.org/Archives/Public/public-xg-emotion/2008Aug/0022.html
Paolo: first issue set/name better discussed in Cannes?
Paolo will distribute spec ahead of Cannes
Enrico will also provide a short overview
Paolo: seperator as white pace easier in XML
Paolo regards this as syntactic issue which may be used consequently
<paolo> xsd:nmtokens
Marc: crucial issue: schema vs. plug in by user
Bjoern asks about inclusion of fusion type for modality as early / late, synchronous / asynchronous
Paolo says that in EMMA there is something like this possibly included
Marc suggests that custom set of modalities would cover for this
Bjoern agrees
Marc favours mode or modality over medium, questioning its necessity.
Marc suggests short list of modalities as a default list, because longer lists are difficult to agree upon quickly.
Marc reminds group of earlier agreement to aim for consistency with EMMA and other formats where possible, but to include only what we need, and to be consistent also within the EmotionML.
Paolo suggests to stay as far as possible consistent with other definitions
Marc asks if there are voices in favour of "medium"
Marc feels that this might be too fine granular for the EMOXL case
Bjoern misses physiologic in basic modalities and context (e.g. driving pattern in a car)
Paolo says from mode medium may be derived
Marc also wants to avoid complexity unless needed
Paolo would like to have opinion of the group hereon
Marc asks on preferences wrt. composite as face&voice
Enrico suggests a survey on this
<marc> http://lists.w3.org/Archives/Public/public-xg-emotion/2008Aug/0001.html
No disagreement on necassity of confidences on core 2,3,4,5,7
Marc asks about sense of confidence for core 1,6,8 and meta 2
Bjoern claims potential importance for core 6 (complex emotions)
Nobody pushing core 1,8 and meta 2
Marc: Felix had suggested confidence for global meta
ok by everybody
Bjoern asks about tag for measurement by which confidence was obtained
Marc: let's note this down as an issue.
<marc> internal progress report: http://lists.w3.org/Archives/Public/public-xg-emotion/2008Sep/0012.html
Marc: weak agremment on global meta data and links to the rest of the world
Marc suggests to focus on these two issues in Cannes and only briefly review rest
everybody agrees.
Bjoern suggests not to make decisions in Cannes but prepare "force" by polls/surveys
Marc: people are invited to give comments
before meeting
... Main topics in Cannes: global meta data, links to the rest of the world
and proposal by Enrico wrt. modalities