Evaluation of DBMS Engines

This is a follow up on a previous post where I posted my criteria for selecting a client/server DBMS. Here I've listed some questions I've come up with during evaluation of DBMS engines. Any insight people could lend to these would be greatly appreaciated. Please respond by email and I will summarize and re-post.
-- Thank you!

START OF QUESTIONS Does the absence of declarative referential integrity in eg. Sybase and SQLBASE (and the resulting need to use triggers to enforce RI) lead to difficulty in implementing and/or maintaining RI? What CASE tools (e.g. Bachman) are available to assist in this?

How easy is it in practice to have a table column that has auto-sequenced numbers (e.g. part numbers). Oracle and Rdb support such a feature. Do the other engines have a workaround for doing this? It is ok if some numbers are missing, as long as the numbers are unique within a table (i.e. 1, 3, 5, 6, 9, is ok).

How important are backward scrolling cursors? (as not all engines support them).

How useful is the ability to retrieve more than one column in a subquery? (Oracle supports this, Sybase does not)

In practice, how much performance is gained by using stored procedures to minimize network traffic?

Thanks in advance for all replys.

• Brian Westrich

--

Brian Westrich

Nutrition Coordinating Center (NCC)     internet: westrich_at_keystone.ncc.umn.edu
University of Minnesota                 phone:    (612) 627 4862