Maybe I shouldn't have mentioned that there is a difference between an information model and a conceptual model, because the kind of conceptual model I'm referring to may not be the one that you think about :-)

This is what I'm talking about:
A conceptual model communicates the meaning of a concept and the characteristics that differentiate a concept from another.

An information model communicates which information about a concept we are interested in for a specific environment.

As an example we could use a lake. In a conceptual model a lake may have the mandatory attribute "deep", because that is a characteristics that can be used for distinguishing a lake from other concepts, but a lake doesn't need to have a name.

In an information model (for a specific environment, e.g. an organization) we may not be interested in how deep a lake is, but the name of the lake is a mandatory attribute because we want to be able to distinguish a specific lake from another lake.

Now, I'm not happy about calling the first model "a conceptual model" because that name is used in many other cicumstances and for other kinds of models. I just haven't met anyone who has been able to suggest a better name for that kind of model, at least not in English :-)


// Dennis Djenfer


On 2010-07-04 15:40, Michael Poulin wrote:


And conceptual model is...

- Michael

------------------------------------------------------------------------
*From:* Steve Jones <[email protected]>
*To:* [email protected]
*Sent:* Fri, July 2, 2010 6:45:55 AM
*Subject:* Re: [service-orientated-architecture] About canonical data model

+1



On 1 July 2010 14:28, Dennis Djenfer <d...@algonet. se <mailto:[email protected]>> wrote:

    Unfortunately we don't have common names for all those different
    kinds of models that we're using in this industry, so it can be
    quite hard sometimes to understand what kind of model someone is
    talking about. I'm talking about a high level information model
    accompanied with a definition and description of each object in
    the model. You may use an ER-tool to create the model, and it's
    only the most important enterprise information objects that goes
    into this model and only the most important relations between
    these objects. The objects may have attributes, but it's only the
    most relevant and important attributes that are used in the model.
    However, it's not a conceptual model, it's an information model.

    // Dennis Djenfer




    On 2010-06-28 22:56, Michael Poulin wrote:
    Dennis,
    if you mean a common (for the enterprise) data vocabulary (what
    is what and how it relates to others), I am with you.

    - Michael

    ------------------------------------------------------------------------
    *From:* Dennis Djenfer <d...@algonet. se> <mailto:[email protected]>
    *To:* service-orientated- architecture@ yahoogroups. com
    <mailto:[email protected]>
    *Sent:* Tue, June 22, 2010 11:05:47 PM
    *Subject:* Re: [service-orientated -architecture] About canonical
    data model

    I agree with that. Data models can be defined on many abstraction
    levels, and when people are talking about using a common data
    model they often mean very different things. I believe in using a
    common enterprise data model as one of the cornerstones when
    integrating systems or creating services, but I'm not talking
    about a detailed data model, rather something like your MDM
    approach or even higher abstraction levels. I've found this high
    level common enterprise data model being very effective at
    creating a common understanding of important concepts and
    information in an enterprise.

    // Dennis Djenfer


    Sounds like the

    On 2010-06-22 22:04, Steve Jones wrote:

    Not really as MDM isn't about the full model its about the core
    and identifying duplicates and commonalities.  So for "Person"
    for instance it might only be 25 attributes that are used by MDM
    to do that.  The other 300+ attributes that exist around the
    enterprise aren't included within the MDM model.

    Steve


    On 23 June 2010 01:37, Dennis Djenfer <d...@algonet. se
    <mailto:[email protected]>> wrote:

        Steve,

        In your blog you write:

        "The only sensible policy is to look at an "active" MDM
        strategy and a brokerage approach to communication between
        systems ideally based around a federated data strategy that
        leaves information its its source systems but provides
        references between them."

        Where does the data model for the active MDM strategy and
        brokerage approach come from? Isn't that the same model as
        you would use for buidling a canonical data model?


        // Dennis Djenfer




        On 2010-06-22 11:34, Steve Jones wrote:
        Short answer... don't.

        Longer answer http://service- architecture. blogspot.
        com/search? q=SOA+canonical
        <http://service-architecture.blogspot.com/search?q=SOA+canonical>

        On 21 June 2010 13:26, <jorg...@uci. cu
        <mailto:[email protected]>> wrote:

            Hello all.

            I have a question for the design of a canonical data model.

            The issue is that I want to create a data services
            layer, and for making the design of services that
            expose data, first I want to create the canonical data
            model, which allows me to properly design services, and
            I want to know if this canonical data model should
            correspond to the data model of the database, E / R
            model , or I just model the information concepts that
            are handled in the database.

            Jorge.









Reply via email to