Loading ...
Sorry, an error occurred while loading the content.

Unified information model

Expand Messages
  • bill.mccrosky
    I may be either asking a question previously discussed in this forum or asking one that belongs in another forum. Please refer me as appropriate ... thanks !
    Message 1 of 11 , Aug 2, 2007
    • 0 Attachment
      I may be either asking a question previously discussed in this forum
      or asking one that belongs in another forum. Please refer me as
      appropriate ... thanks !

      I am on a team that is "modernizing" a 20 year old application. It is
      a cobol-cics-vsam application; we are migrating it to a
      java-websphere-db2 application. Actually the target architecture is
      more complex (WTX, WPS, etc), but I want to stay focused on my
      concerns about information.

      I'd like to have a "unified" understanding of the information in the
      system, and I want to control/audit this understanding without being
      bureaucratic or otherwise slowing down development.

      I am concerned that object models, message models, and data(base)
      models will evolve with less-than-perfect integration of the
      information semantics - the object modelers will build their classes,
      services and components will exchange messages, and the database will
      be built - all without enough thought being given to whether objects
      define data consistently, exchange messages with data having the same
      meaning at the target as at the source, and yet another meaning in the
      database.

      How do I get control of all of this information without slowing down
      development? What methodologies, functionality/tools should I be
      looking at? Thanks.
    • Curt Sampson
      ... This may be too big a change for your group to handle, but you might look at Extreme Programming. It focuses not on planning everything perfectly, but
      Message 2 of 11 , Aug 2, 2007
      • 0 Attachment
        On Thu, 2 Aug 2007, bill.mccrosky wrote:

        > How do I get control of all of this information without slowing down
        > development? What methodologies, functionality/tools should I be
        > looking at? Thanks.

        This may be too big a change for your group to handle, but you might
        look at Extreme Programming. It focuses not on planning everything
        perfectly, but maximizing communication between everyone on the project
        and having the ability to fix things very, very quickly when an issue is
        discovered.

        If you've not done it before, hiring a coach is highly recommended.
        The only company that does coaching that I've worked with (besides my
        own) is Industrial Logic, but they're pretty good. And I'd expect that
        ThoughtWorks would also be good, given the calibre of staff they have.

        cjs
        --
        Curt Sampson <cjs@...> +81 90 7737 2974
        http://www.starling-software.com
        The power of accurate observation is commonly called cynicism
        by those who have not got it. --George Bernard Shaw
      • Scott Ambler
        A few thoughts which may or may not help: 1. The real issue isn t getting control over the effort, it s motivating people to do the right things. 2. Can you
        Message 3 of 11 , Aug 3, 2007
        • 0 Attachment
          A few thoughts which may or may not help:
          1. The real issue isn't getting control over the
          effort, it's motivating people to do the right things.
          2. Can you educate developers on the importance of
          these issues?
          3. Can you provide "support services" to the
          developers to help them to do these things which
          reflects the way that they work? What minimal amount
          of work would they need to do to get the job done?

          - Scott
          --- "bill.mccrosky" <mccrosky@...> wrote:

          > I may be either asking a question previously
          > discussed in this forum
          > or asking one that belongs in another forum. Please
          > refer me as
          > appropriate ... thanks !
          >
          > I am on a team that is "modernizing" a 20 year old
          > application. It is
          > a cobol-cics-vsam application; we are migrating it
          > to a
          > java-websphere-db2 application. Actually the target
          > architecture is
          > more complex (WTX, WPS, etc), but I want to stay
          > focused on my
          > concerns about information.
          >
          > I'd like to have a "unified" understanding of the
          > information in the
          > system, and I want to control/audit this
          > understanding without being
          > bureaucratic or otherwise slowing down development.
          >
          >
          > I am concerned that object models, message models,
          > and data(base)
          > models will evolve with less-than-perfect
          > integration of the
          > information semantics - the object modelers will
          > build their classes,
          > services and components will exchange messages, and
          > the database will
          > be built - all without enough thought being given to
          > whether objects
          > define data consistently, exchange messages with
          > data having the same
          > meaning at the target as at the source, and yet
          > another meaning in the
          > database.
          >
          > How do I get control of all of this information
          > without slowing down
          > development? What methodologies,
          > functionality/tools should I be
          > looking at? Thanks.
          >
          >
          >

          Scott W. Ambler
          Practice Leader Agile Development, IBM Methods Group
          http://www-306.ibm.com/software/rational/bios/ambler.html


          Get news delivered with the All new Yahoo! Mail. Enjoy RSS feeds right on your Mail page. Start today at http://mrd.mail.yahoo.com/try_beta?.intl=ca
        • bill.mccrosky
          Fortunately, at this stage of the game, almost all things are possible. We have a high-level (subject to iterative refinement) solution architecture of the
          Message 4 of 11 , Aug 3, 2007
          • 0 Attachment
            Fortunately, at this stage of the game, almost all things are
            possible. We have a high-level (subject to iterative refinement)
            solution architecture of the business processes, components services.
            My next step is to identify the information services to support this
            solution architecture.

            You're probably right that obsession with control (my words, not
            yours) will be counterproductive, but who controls the semantics of
            the information? Who decides - whether by command-and-control or by
            collaborative empowerment of the developers - on the shared
            understanding of the "information model"? I want the developers of
            the business processes to agree on what - for example - the
            createCustomer, getCustomer, or deleteCustomer information services do
            with customer-related information. I also think - correct me if I am
            wrong - that all developers share a common object model. Element
            names should be the same, attributes should be the same, etc. I want
            developers to be able to cross-train on the services, right? The same
            goes for messages being passed between objects. I don't want two
            developers to suboptimally decide that THEIR message is strucutred one
            way and then another developer develops his own variation of that message.

            We ARE following industry standards, supplemented by (IBM) reference
            models, which will help standardize object models, data models, and
            message models. Is that a sufficient platform on which to let a
            thousand collaborative flowers bloom?
            --- In agileDatabases@yahoogroups.com, Scott Ambler <scottwambler@...>
            wrote:
            >
            > A few thoughts which may or may not help:
            > 1. The real issue isn't getting control over the
            > effort, it's motivating people to do the right things.
            > 2. Can you educate developers on the importance of
            > these issues?
            > 3. Can you provide "support services" to the
            > developers to help them to do these things which
            > reflects the way that they work? What minimal amount
            > of work would they need to do to get the job done?
            >
            > - Scott
            > --- "bill.mccrosky" <mccrosky@...> wrote:
            >
            > > I may be either asking a question previously
            > > discussed in this forum
            > > or asking one that belongs in another forum. Please
            > > refer me as
            > > appropriate ... thanks !
            > >
            > > I am on a team that is "modernizing" a 20 year old
            > > application. It is
            > > a cobol-cics-vsam application; we are migrating it
            > > to a
            > > java-websphere-db2 application. Actually the target
            > > architecture is
            > > more complex (WTX, WPS, etc), but I want to stay
            > > focused on my
            > > concerns about information.
            > >
            > > I'd like to have a "unified" understanding of the
            > > information in the
            > > system, and I want to control/audit this
            > > understanding without being
            > > bureaucratic or otherwise slowing down development.
            > >
            > >
            > > I am concerned that object models, message models,
            > > and data(base)
            > > models will evolve with less-than-perfect
            > > integration of the
            > > information semantics - the object modelers will
            > > build their classes,
            > > services and components will exchange messages, and
            > > the database will
            > > be built - all without enough thought being given to
            > > whether objects
            > > define data consistently, exchange messages with
            > > data having the same
            > > meaning at the target as at the source, and yet
            > > another meaning in the
            > > database.
            > >
            > > How do I get control of all of this information
            > > without slowing down
            > > development? What methodologies,
            > > functionality/tools should I be
            > > looking at? Thanks.
            > >
            > >
            > >
            >
            > Scott W. Ambler
            > Practice Leader Agile Development, IBM Methods Group
            > http://www-306.ibm.com/software/rational/bios/ambler.html
            >
            >
            > Get news delivered with the All new Yahoo! Mail. Enjoy RSS
            feeds right on your Mail page. Start today at
            http://mrd.mail.yahoo.com/try_beta?.intl=ca
            >
          • bill.mccrosky
            ... Yes I m looking at various extreme/agile approaches to the project. My management is skeptical but is considering the possibility. I am going to take a
            Message 5 of 11 , Aug 3, 2007
            • 0 Attachment
              --- In agileDatabases@yahoogroups.com, Curt Sampson <yahoo@...> wrote:
              > This may be too big a change for your group to handle, but you might
              > look at Extreme Programming.

              Yes I'm looking at various extreme/agile approaches to the project.
              My management is skeptical but is considering the possibility. I am
              going to take a scrum master course.
            • Rick Wood
              Here s a related question (which may also already have been answered). Let s say you want to generate test data for a system. There are two datasets. The first
              Message 6 of 11 , Aug 3, 2007
              • 0 Attachment
                Here's a related question (which may also already have been answered).
                Let's say you want to generate test data for a system. There are two
                datasets. The first is inserted into the database, so you can test
                performance, and test the components that read data from the db. The
                second dataset is presented to the front end, so you can test its
                transformation by business logic.

                Can these two datasets be generated from the same model?

                Rick

                bill.mccrosky wrote:
                >
                > I may be either asking a question previously discussed in this forum
                > or asking one that belongs in another forum. Please refer me as
                > appropriate ... thanks !
                >
                > I am on a team that is "modernizing" a 20 year old application. It is
                > a cobol-cics-vsam application; we are migrating it to a
                > java-websphere-db2 application. Actually the target architecture is
                > more complex (WTX, WPS, etc), but I want to stay focused on my
                > concerns about information.
                >
                > I'd like to have a "unified" understanding of the information in the
                > system, and I want to control/audit this understanding without being
                > bureaucratic or otherwise slowing down development.
                >
                > I am concerned that object models, message models, and data(base)
                > models will evolve with less-than-perfect integration of the
                > information semantics - the object modelers will build their classes,
                > services and components will exchange messages, and the database will
                > be built - all without enough thought being given to whether objects
                > define data consistently, exchange messages with data having the same
                > meaning at the target as at the source, and yet another meaning in the
                > database.
                >
                > How do I get control of all of this information without slowing down
                > development? What methodologies, functionality/tools should I be
                > looking at? Thanks.
                >
                >
              • Curt Sampson
                ... The developers themselves. You can t force someone to understand something the way you do; you can only talk them and come to some sort of agreement,
                Message 7 of 11 , Aug 4, 2007
                • 0 Attachment
                  On Fri, 3 Aug 2007, bill.mccrosky wrote:

                  > Who decides - whether by command-and-control or by collaborative
                  > empowerment of the developers - on the shared understanding of the
                  > "information model"?

                  The developers themselves. You can't force someone to understand
                  something the way you do; you can only talk them and come to some sort
                  of agreement, assuming you're both willing.

                  I'd suggest you start with a small group of developers, only your best
                  and most experienced designers. Spend a copule of weeks getting the
                  basic outline of the application designed and coded, and then start
                  adding other developers. Pair each new developer with an existing one,
                  at least for a week or so anyway, so that they can learn the design and
                  the culture.

                  > I don't want two developers to suboptimally decide that THEIR message
                  > is strucutred one way and then another developer develops his own
                  > variation of that message.

                  If you set up a comprehensive automated test suite, it's easy to go back
                  and fix this kind of stuff when you discover it. Easier, in fact, than
                  spending a lot of time planning in order to try to avoid these things
                  happening.

                  cjs
                  --
                  Curt Sampson <cjs@...> +81 90 7737 2974
                  http://www.starling-software.com
                  The power of accurate observation is commonly called cynicism
                  by those who have not got it. --George Bernard Shaw
                • Scott Ambler
                  ... My first reaction is Sure, why not? . You re working on the same system on the same team so the same conceptual information model (or whatever you want
                  Message 8 of 11 , Aug 7, 2007
                  • 0 Attachment
                    --- Rick Wood <rwood@...> wrote:

                    > Here's a related question (which may also already
                    > have been answered).
                    > Let's say you want to generate test data for a
                    > system. There are two
                    > datasets. The first is inserted into the database,
                    > so you can test
                    > performance, and test the components that read data
                    > from the db. The
                    > second dataset is presented to the front end, so you
                    > can test its
                    > transformation by business logic.

                    My first reaction is "Sure, why not?". You're working
                    on the same system on the same team so the same
                    conceptual information model (or whatever you want to
                    call it) is applicable. So, why can't this model
                    drive the test data for both types of test? As long
                    as you're putting whatever it is that you're testing
                    into a known state before running the tests you should
                    be OK.

                    My second reaction is that it could be tough for some
                    organizations to do this because the developers and
                    the data folks aren't working together on the same
                    team. This to me is a serious mistake.

                    - Scott


                    Scott W. Ambler
                    Practice Leader Agile Development, IBM Methods Group
                    http://www-306.ibm.com/software/rational/bios/ambler.html


                    Be smarter than spam. See how smart SpamGuard is at giving junk email the boot with the All-new Yahoo! Mail at http://mrd.mail.yahoo.com/try_beta?.intl=ca
                  • Rick Wood
                    I agree they could come from the same conceptual model. What interests me about generated data is, can it say anything about how good the model is? In a sense,
                    Message 9 of 11 , Aug 7, 2007
                    • 0 Attachment
                      I agree they could come from the same conceptual model. What interests
                      me about generated data is, can it say anything about how good the model
                      is? In a sense, generating data from a model is the inverse of the
                      analysis that created the model from the data. If generated data can't
                      replace real data, is that a problem with the generation, or is
                      something missing in the model itself? This is also pretty much true of
                      code generation from a business object model.

                      But I have not seen much activity around generating data from a data
                      model. Is there any work going on there?

                      - Rick

                      Scott Ambler wrote:
                      >
                      >
                      > --- Rick Wood <rwood@... <mailto:rwood%40siteduck.com>> wrote:
                      >
                      > > Here's a related question (which may also already
                      > > have been answered).
                      > > Let's say you want to generate test data for a
                      > > system. There are two
                      > > datasets. The first is inserted into the database,
                      > > so you can test
                      > > performance, and test the components that read data
                      > > from the db. The
                      > > second dataset is presented to the front end, so you
                      > > can test its
                      > > transformation by business logic.
                      >
                      > My first reaction is "Sure, why not?". You're working
                      > on the same system on the same team so the same
                      > conceptual information model (or whatever you want to
                      > call it) is applicable. So, why can't this model
                      > drive the test data for both types of test? As long
                      > as you're putting whatever it is that you're testing
                      > into a known state before running the tests you should
                      > be OK.
                      >
                      > My second reaction is that it could be tough for some
                      > organizations to do this because the developers and
                      > the data folks aren't working together on the same
                      > team. This to me is a serious mistake.
                      >
                      > - Scott
                      >
                      > Scott W. Ambler
                      > Practice Leader Agile Development, IBM Methods Group
                      > http://www-306.ibm.com/software/rational/bios/ambler.html
                      > <http://www-306.ibm.com/software/rational/bios/ambler.html>
                      >
                      > Be smarter than spam. See how smart SpamGuard is at giving junk email
                      > the boot with the All-new Yahoo! Mail at
                      > http://mrd.mail.yahoo.com/try_beta?.intl=ca
                      > <http://mrd.mail.yahoo.com/try_beta?.intl=ca>
                      >
                    • Scott Ambler
                      I don t know off hand about anyone generating test data from a data model but don t see why it couldn t be done. Might not be the prettiest data in the world
                      Message 10 of 11 , Aug 7, 2007
                      • 0 Attachment
                        I don't know off hand about anyone generating test
                        data from a data model but don't see why it couldn't
                        be done. Might not be the prettiest data in the world
                        though...

                        - Scott
                        --- Rick Wood <rwood@...> wrote:

                        > I agree they could come from the same conceptual
                        > model. What interests
                        > me about generated data is, can it say anything
                        > about how good the model
                        > is? In a sense, generating data from a model is the
                        > inverse of the
                        > analysis that created the model from the data. If
                        > generated data can't
                        > replace real data, is that a problem with the
                        > generation, or is
                        > something missing in the model itself? This is also
                        > pretty much true of
                        > code generation from a business object model.
                        >
                        > But I have not seen much activity around generating
                        > data from a data
                        > model. Is there any work going on there?
                        >
                        > - Rick


                        Scott W. Ambler
                        Practice Leader Agile Development, IBM Methods Group
                        http://www-306.ibm.com/software/rational/bios/ambler.html


                        Ask a question on any topic and get answers from real people. Go to Yahoo! Answers and share what you know at http://ca.answers.yahoo.com
                      • Markus Samuel Gallagher
                        The Data Dude in the Microsoft Visual Studio product line supports generating test data from an existing Schema. As Scott already mentioned... the data isn t
                        Message 11 of 11 , Aug 19, 2007
                        • 0 Attachment
                          The Data Dude in the Microsoft Visual Studio product line supports
                          generating test data from an existing Schema. As Scott already
                          mentioned... the data isn't the prettiest... however there are
                          possibilities to generate according to regex or via lookups in other
                          tables, as well as coding custom data generation extensions.

                          --- In agileDatabases@yahoogroups.com, Scott Ambler <scottwambler@...>
                          wrote:
                          >
                          > I don't know off hand about anyone generating test
                          > data from a data model but don't see why it couldn't
                          > be done. Might not be the prettiest data in the world
                          > though...
                          >
                          > - Scott
                          > --- Rick Wood <rwood@...> wrote:
                          >
                          > > I agree they could come from the same conceptual
                          > > model. What interests
                          > > me about generated data is, can it say anything
                          > > about how good the model
                          > > is? In a sense, generating data from a model is the
                          > > inverse of the
                          > > analysis that created the model from the data. If
                          > > generated data can't
                          > > replace real data, is that a problem with the
                          > > generation, or is
                          > > something missing in the model itself? This is also
                          > > pretty much true of
                          > > code generation from a business object model.
                          > >
                          > > But I have not seen much activity around generating
                          > > data from a data
                          > > model. Is there any work going on there?
                          > >
                          > > - Rick
                          >
                          >
                          > Scott W. Ambler
                          > Practice Leader Agile Development, IBM Methods Group
                          > http://www-306.ibm.com/software/rational/bios/ambler.html
                          >
                          >
                          > Ask a question on any topic and get answers from real people.
                          Go to Yahoo! Answers and share what you know at
                          http://ca.answers.yahoo.com
                          >
                        Your message has been successfully submitted and would be delivered to recipients shortly.