Are We Working On The Right Problems?

Plenary Panel
Michael Stonebraker
Informix Corporation
mike@postgres.Berkeley.eduU

 Abstract
There appears to be a discrepancy between the research topics being pursued by the database research community and the key problems facing information systems decisions makers such as Chief Information Officers (CIOs). Panelists will present their view of the key problems that would benefit from a research focus in the database research community and will discuss perceived discrepancies. Based on personal experience, the most commonly discussed information systems problems facing CIOs today include:

1. Packaged Applications
Everybody is doing "buy", not "build". That completely insulates users from the DBMS. Hence, they don't care what DBMS gets used. All packaged applications are "agnostic", i.e. run on everybody's DBMS. Will this make our research area irrelevant?

2. Legacy Systems
The "smokestack" systems from 20 years ago are the Achilles' heel of everybody. They seem to drive everything. There seem to be very few good ideas of what to do in this area. So what are we doing here?

3. Middleware
Should it be messaging, application server, or DBMS? How can one choose between CORBA, DCOM, and RMI given the rift between Microsoft and the rest of the world on these issues? So what are we doing about this?

 4. Enterprise Data Integration
Enterprise data integration is the top item on every CIOs wish list. So what are we doing about it?

5. "No Knobs"
DBMSs (especially one of the popular ones) are way too hard to install, setup, tune, and maintain. When are we going to learn something from the vendors of children's games and get real here?

6. User Interfaces
The DBMS research community is out to lunch here. They have contributed nearly nothing to this important area. Program committee after program committee routinely bows to this alter and then rejects the submitted papers. Are we simply abdicating this space?

7. Software Productivity
Hardware productivity goes up by a factor of two every two years. Software productivity has gone up less than a factor of two in the last 20 years. What can be done about this?

8. IT Project Failure
It is estimated that 70-80% of all IT projects fail. If civil engineers built bridges like we build software, modern society would fire them. What right do we have to call ourselves professionals? What can be done here?