Recommendations

  • (1) DON'T TRY TO CREATE A PERFECT MODEL IN THE FIRST RUN!

  • (2) START TESTING AS EARLY AS POSSIBLE (or even earlier)

  • (3) PLAN FOR SEVERAL CYCLES OF REFINEMENT

  • (4) WHAT EVER THE RESULT IS – CHANGE WILL HAPPEN

____________________________________________________________________________________

The information model defines all the terms and structures for the development of detailed process specifications and finally for the development of a system.

Business requirements tend to be vague and ambiguous if they are not based on a properly documented information model.

Quality Assure Data Models:

  • (1) Get (or create) real life data to test the model.
  • (2) Store the data in formally structured Excel-Sheets, i.e. one table for each entity with columns containing the values of the attributes
  • (3) Load the Excel-Sheets into the repository -> MetaEXL facility of metasafe
  • (4) Define UseCases and check the data against these UseCases
  • (5) Use the MetaEditor and especially the erSQL queries to extract the data which are required for the Use Cases.
  • (6) If the model does not fit return to the modeler and update the model.