Oracle CASE Tools

I am looking for CASE tool support for a large-scale enterprise re-engineering project. The current disjoint stovepipe systems use both Oracle 6 and Sybase. We will need to reengineer a new logical data model from the existing Oracle and Sybase DDL. We will need to examine the data models of the current disjoint applications, look for areas of overlap, and develop a new global model that can accommodate the needs of the various applications. I am expecting the new global model to have on the order of 300-500 entities.

I would like to consider Oracle's CASE tools as a candidate for this project. We have never used them before, so I am not familiar with their good points, bad points, etc. Having been burned by CASE tools before, I am reluctant to make a recommendation whether to buy Oracle CASE, even just for evaluation purposes, without getting some feedback on the tools from real experienced users. Hence this posting. I would like to hear from experienced Oracle CASE users as to their experiences with Oracle CASE on large-scale projects.

Some specific questions:

1. Have you developed logical schemas that are as large as 300-500 entities? If so, was the model understandable or was it just so much undecipherable spaghetti? Do the CASE tools provide support for looking at the model at various levels of abstractions so that different individuals can see the model at different levels of details?
2. How good is Oracle's reverse engineering support? Can it do a reasonable job of reverse engineering a couple of hundred DEFINE TABLE statements? Does it do a reasonable job of identifying relationships between entities? Are there any gotchas I should watch out for? Will it support reengineering of Sybase DDL? I've seen some recent postings that indicate some versions of the tools do not reverse engineer column definitions.
3. Is there any support for identifying and resolving ambiguities between the current disjoint data models? For example, if department A and department B both have individual systems today, both have information on employee's addresses, but have the address information in different formats, is there any support in the CASE tools for identifying this and coming up with a schema that will satisfy both departments information needs?
4. Is there any support for maintaining traceability between old and new data models? For example, can I indicate that a section of the new logical model is based on a section of Department A's old schema?
5. The current applications are a mixture of SQL*Forms 2.3, 3.x, and just about any other kind of code with embedded SQL that you can possibly imagine. The client's stated goal is to move away from SQL*Forms. If we aren't going to use the generators to produce SQL*FOrms applications, are the CASE tools still worthwhile? We will also probably still have Sybase in the system. Does this compromise the usefulness of the Oracle tools?
6. Have you ever used Oracle's tools on a government program that invoked DOD-STD-2167A? If so, what kind of problems did you have using Oracle's tools to produce 2167A-compliant documentation?
7. Are there any other issues that you think I should be aware of in looking for CASE tools? Are there any other CASE tools you think should be considered as candidates?

Thanks in advance,

Richard Howey
prefered email: rhowey_at_mhs.sp.paramax.com Received on Wed Jul 27 1994 - 19:04:39 CEST