See also: IRC log
1. It was agreed that the questionnaire about requirement priorities would be used as a starting point for a list of high-priority items. Ian will compile a list of items above the (arbitrary) threshold of 70% "must have" or "should have" answers. This list can then be commented, group participants can discuss whether items should be added or removed from the "high priority" list. After agreement, this list should be the basis for the first draft specification.
2. Regardingpossible syntactic realisations of an Emotion Markup Language, it was agreed that the language should encode information explicitly, even if that may mean that simple cases will not look simple. Interoperability was highlighted as a key issue, and as most of the relevant languages are in XML, that may push towards XML. Some group members felt that it was not appropriate to try and define an ontology of emotions; this may be too concrete, not in line with the aim of the group to create a language which is neutral with respect to concrete emotion theories. This may be a reason for not preferring OWL.
As one key requirement, the need for "plugging in" a custom set of descriptors was discussed in more details; for example, a user should be able to indicate a custom set of emotion categories, ideally directly within a document, without having to define a new "dialect" of the emotion markup. It may be that a concept of "sub-classing" a generic class with a more specific class would be helpful for this. The situation with respect to "sub-classing" in various formalisms requires clarification, which will be sought before the next meeting.
Two summaries of the results:
<marc> http://lists.w3.org/Archives/Public/public-xg-emotion/2008Jan/0015.html
<marc> http://lists.w3.org/Archives/Public/public-xg-emotion/2008Jan/0016.html
marc: says feedback for questionairre was
good
... notes people from outside the group gave valuable feedback
... notes that results of questionairre are in general agreement
... notes that emotion categories and dimensions are agreed must haves
ian:perhaps we say over 75% are definately in?
<bjoern> or over 70% (3 important ones @ 71%...)
ian: mentions that some items may be easy to impliment through the structure of XML
marc: these may require a detailed discussion that may be too in depth for now
<marc> a suggestion: we could draw a line at, e.g., 70%,
<marc> and suggest that those above 70% are "definitely in".
<marc> That gives us a list of 11 requirements,
<marc> and then people can discuss why the list should be changed.
<bjoern> agrees.
ian:as an aside i will split the list into "meta data", "body data" and "link data"
<scribe> ACTION: ian to create list of "first cut" items to discuss and circulate due by tomorrow [recorded in http://www.w3.org/2008/01/31-emotion-minutes.html#action01]
<trackbot-ng> Created ACTION-9 - Create list of \"first cut\" items to discuss and circulate due by tomorrow [on Ian Wilson - due 2008-02-07].
dylan: asks about confidence ratings, is there one meta or confidence ratings for each item
marc: we will discuss that later when we address each item
<marc> owl example from bill
<marc> rdf example from ian
<marc> deep xml example from myriam
<marc> shallow xml example from bjoern
<marc> one key choice: how explicit?
<marc> beiing explicit would mean giving up on "simple cases should look simple"
ian: annotations and models have somewhat differing requirements
<Myriam> I thing it's better to use simple cases in the first step
marc: we want to stay with 1 format however
<Myriam> Deep structure is more flexible
marc: do we want to trade simplicity for explicitness?
<marc> is there an agreement that we want to be explicit?
<Myriam> yes
<bjoern> yes
ian:yes
<bjoern> yes
ian: how important is it that we can interoperate with other formats and other groups?
dylan: agrees, that interoperation is crucial
<Myriam> the mapping will be more or less complex
dylan: thinks going to an ontology is a step too far and making a comitment to an ontology is too much
ian: agrees, thinks an ontology would be too concrete wheras xml is more abstract
<Myriam> perhaps, RDF is an intermediate between XML and OWL
<Myriam> it's possible to use like XML
<Myriam> but also to create an ontology
<Myriam> XML or RDF namespace are the same goal
<Myriam> the disambiguous of type
<Myriam> but some RDF namespaces are a norm
<Myriam> like DC
ian: rdf requires unambiguous namespaces and terms
<Myriam> yes
<Myriam> and XML accept qualified or unqualified terms
<Myriam> but we can force the utilisation of namespace
marc: wants to know if xml, rdf and owl can be sub classed?
<scribe> ACTION: ian to investigate sub classing in RDF due next friday [recorded in http://www.w3.org/2008/01/31-emotion-minutes.html#action02]
<trackbot-ng> Created ACTION-10 - Investigate sub classing in RDF due next friday [on Ian Wilson - due 2008-02-07].
<Myriam> I think also for OWL because it's based on object programming
ian:do we have a link for sub classing XML?
<Myriam> If you write a simpleType to restrict xsd:integer
<Myriam> to range [0,1]
<Myriam> we can also use xsd:restruction from a complexType
<Myriam> in XSD
ian:can you find an example?
<Myriam> in http://www.iut.univ-paris8.fr/~lamolle/EMOXG/emoxg-basetype.xsd
<Myriam> we can see basic type for EMOXG
Christian: says we dont want to be too explicit, only with core attributes
<Myriam> and in http://www.iut.univ-paris8.fr/~lamolle/EMOXG/emoXG0.1.xsd
<Myriam> we can see I extend or restrict basic type
<marc> Let's investigate how the various formalisms can deal with the "plug in" challenge
<Myriam> I think it exist 3 formalisms to "plug in"... in XSD -- can we use the same examples to compare the "plug-in"? or the sub-classsing?
<scribe> ACTION: Myriam to investigate subclassing in XML due 2008-02-21 [recorded in http://www.w3.org/2008/01/31-emotion-minutes.html#action03]
<trackbot-ng> Sorry, couldn't find user - Myriam
<scribe> ACTION: Bill to investigate subclassing in OWL due 2008-02-21 [recorded in http://www.w3.org/2008/01/31-emotion-minutes.html#action04]
<trackbot-ng> Created ACTION-11 - Investigate subclassing in OWL due 2008-02-21 [on Bill Jarrold - due 2008-02-07].
<marc> Let's note for the future: another question we will have to look at is how the language can be "plugged into" another language such as EARL or SMIL etc.
<marc> ACTION: Marc to describe the challenge of plugging in different categories etc. within the language - due 15 Feb [recorded in http://www.w3.org/2008/01/31-emotion-minutes.html#action05]
<trackbot-ng> Created ACTION-12 - describe the challenge of plugging in different categories etc. within the language [on Marc Schröder - due 2008-02-15].
<marc> Next meeting should be the "late slot" (15:00 UTC); I suggest the week of 25 February, day to be determined with Doodle.
all: agreed
<bjoern> all:agreed,2
<Dylan> Thanks everyone. Bye!
<bjoern> bye
<Myriam> bye