Thank you so much for that clarification. Unfortunately, it is a devastating problem for us now.
My task is to design and build information systems for a billion dollar corporation where no automation of any type currently exists - no information systems and no applications. The corporation is comprised of 11 very large subsidiaries - all with basically the same common needs - hiring, payroll, purchasing, logistics, etc. So we are in the process of building the enterprise data model which would serve the purposes for all of the affiliate companies. At this time, we have collected enough use cases and created enough ERDs to begin generating the first applications such as an electronic application for employment, personnel management, time keeping, etc., etc. All of which can be re-used (with some differences) by the different companies, and all of which will use many of the same tables such as employee, timesheet, schedules, etc., etc.
We approach the problem in the standard way -
- write the problem statement
- create use cases and capture requirements
- identify the entities/classes and create ERDs
This done, we would now like to generate a class diagram to match the ERD that illustrates the data necessary to implement, for example, an online electronic application for employment. The problem is that VP will not, under any circumstances, give me a class diagram that consists of only the components involved in the solution. Instead, it will only give us the ‘big bang’ class diagram, which is a mess, is not usable as is without hours of work to re-organize it aesthetically to make it usable, then cut and past only the classes we are interested in to a new class diagram that matches the ERD.
If this was not bad enough, then there is the code generation. We can not simply generate only the code we are interested in, we must generate code for every single entity in the internal repository of VP.
The only solution at this point is to create individual projects where the same work - the same entities, definitions, date types and specifications are replicated over and over again with the multitude of individual projects we will be required to created.
What this effectively does it devout most of our time managing a case tool versus addressing the serious problems this company is faced with. The case tool definitely does not emulate how software is actually developed using the proven methods of our industry.
Don’t mean to be critical, but these are not opinions, these are the facts that I am realistically faced with at the moment, and to be honest, I am perplexed as to what to do.
At any rate, thank you for your reply. You help is very much appreciated.
Best Regards,
Ricky Miller
