Re: [Ag Meinungsfindungstool] Common business entity model

[Michael Allan <mike AT zelea.com>]

Marc and George,

Thanks Marc, I have a rough understanding of the model now.  If any
other English speakers have questions about it, then I'll try to
answer them.  Sometimes explaining is the best way to learn.  ;^)

Marc said:
> That's true. But IMHO if a Group is composed of 2..* Participants,
> that doesn't mean a Participant can't be member of several
> Groups. So maybe we need a * multiplicity at the Groups side?

Or maybe an open diamond symbol (aggregate).  I think filled diamond
(composite) implies a single group.


George said:
> So, a couple of simple domain models for argumentative discussions
> and consensus already exist:
> 1. A consensus ontology, based on IBIS, problem marketplaces and
> idea management:
> http://www.imc.com.gr/ontologies/eDialogos/consensus/

Welcome George,

http://www.imc.com.gr/ontologies/eDialogos/consensus/img/class_diagram_v0.1.png
I was looking at the core relations (eDialogos below).  I think I can
map them to the terminology we use (in Outcast) and also AG Meinungs-
findungstool (CBEM/AG):


                  has                has
  IdeaContest  <-------  Idea  ----------->  User          (eDialogos)

         Poll  <-------  Position  ------->  Person        (Outcast)

        Topic  <-------  Opinion  -------->  Participant   (CBEM/AG)


Maybe this can be our Rosetta stone for the rest of the world. :-)

Michael


marc said:
> Michael Allan wrote:
> >Marc said:
> >> Here is a first draft of our idea of a Common Business Entity Model
> >> (where 'Business' means it's not a 'Data' Entity Relationship Model
> >> assignable to a database schema, it's just objects found in the real
> >> world - the database schema might look totally different):
> >> http://wiki.piratenpartei.de/wiki/images/4/45/DSFS-BusinessEntityModel.jpg
> 
> >I have some questions about the model.  To keep them separate from my
> >other questions, I opened this thread.
> 
> >(a) Why is discussion itself not modeled?  It's crucial of course, but
> >I don't see it anywhere.
> 
> This model is currently meant to be the base solely for the discussion 
> system!
> Therefore the discussion itself is a 'Process' or 'Workflow' not an 
> 'Entity'.
> 
> >(b) Why is a vote included as an actual (real world) entity?  Myself,
> >I think of voting as a formality of the system, something new we
> >introduce.  It does have *correlates* in the real world, but people
> >don't actually go about casting votes in order to form their opinions,
> >not yet.
> 
> Ok. Sorry. This is definitely a problem in translation. Please read 
> 'RATING' 
> instead of 'VOTE' here!
> 
> >(In fact, most of the right side seems to be modeling a
> >artificial system yet to be deployed, not the real world.  I probably
> >mistook your meaning for "business entity".)
> 
> Yeah, that's me too ;o) Business Entity is not a good term at all. What it 
> wants to express is, that it is not a DATA ENTITY, that can be mapped 
> directly to a database schema. Even if sometimes it would ;o)
> 
> In fact we are modeling the entities of an artificial system here. But 
> maybe 
> we can use the term 'Domain Entity' as George Anadiotis suggested. That's 
> more close to what I think the entities are all about.
> 
> >Looking at the left side, I see a few possible errors:
> 
> >(c) An opinion is not really (in itself) a document.  A document can
> >be involved in the formalization and expression of an opinion, but the
> >opinion is not bound to that form.
> 
> Yes you are right. So it is not realy about real world entities but more 
> about the domain entities that are reflected by the system. So sorry for 
> misleading you!
> 
> In our working group we decided, that the Discussion System is only able to 
> deal with textual input. No video or audio processing capabilities. That's 
> why at the end everything boils down to some combination of characters.
> 
> >(d) Likewise, a point of view is not an opinion (at least not in
> >English).  It is one's outlook from a personal situation.  It may
> >influence opinion in peculiar ways, but is not an opinion in itself.
> 
> Ok. That's unfortunately also lost in translation I believe. Please read 
> POSITION instead of POINT OF VIEW. The groups position is a consensus of 
> all 
> individual opinions of the group members.
> 
> >(e) Group membership is not a closed composite relation, but rather an
> >open aggregate.  One can be a member of many groups.
> 
> That's true. But IMHO if a Group is composed of 2..* Participants, that 
> doesn't mean a Participant can't be member of several Groups. So maybe we 
> need a * multiplicity at the Groups side?
> 
> >Although I don't have a full understanding yet, I suspect it will be
> >difficult to find the right model to code by.  It may take a while.
> >Later (in the other thread), I'll try to suggest a way of buying time
> >and (possibly) evolving some of the models by discovery.
> 
> Ok. Thank you!
> 
> Cheers
> Marc 


George Anadiotis said:
> Dear Marc and Michael, all
> 
> i hope i am not "interrupting" your discussion, but it seems to me there is
> something here which i may be able to contribute - even though
> unfortunately i have not been active on the list, i still do some work on
> the topic.
> 
> So, i am glad to see that the discussion is getting active and entering
> what appears to me like a standardization / integration stage. Typically,
> there are 2 ways to deal with interoperability - on the functionality and
> on the data level. So in your example, that would translate to having T1
> being able to call some API on T2 in order to perform functions on it on
> behalf of a T1 user, and vice versa, and having T1 and T2 exchange some
> data in order to process them.
> 
> Some common denominator can be agreed on on both levels, however i would
> like to emphasize more on the data side of things here. It seems to me that
> what you are doing is trying to come up with a schema to describe basic
> domain entities, their properties and relations - in other words, come up
> with a domain ontology. I do not want to lecture anyone here or to go on
> extensively about this, but i would just like to point out to a body of
> scientific work on the subject that you may be interested in - there's no
> point in reinventing the wheel after all.
> 
> So, a couple of simple domain models for argumentative discussions and
> consensus already exist:
> 1. A consensus ontology, based on IBIS, problem marketplaces and idea
> management: http://www.imc.com.gr/ontologies/eDialogos/consensus/
> 2. An argumentative discussion ontology, based on
> IBIS<http://en.wikipedia.org/wiki/Issue-Based_Information_System>and
> SIOC <http://ceur-ws.org/Vol-494/coinpaper8.pdf>: http://
> ceur-ws.org/Vol-405/paper4.pdf
> 3. A survey of existing work in the domain:
> http://www.semantic-web-journal.net/sites/default/files/swj138_2.pdf
> 
> The 1st and 2nd approach can be harmonized, and this is set as future work.
> There is also a body of work on the aspect of consensus on the ontology
> development process itself which could be interesting here. Hope this
> helps, let me know if you are interested in more details.
> 
> best regards
> George Anadiotis