I remember in the old DOS days, there were a lot of database designs with one table representing more than one data entity. A very simple example being one table that holds data for vendors, customers, contacts, etc. I've never designed databases this way but I though there would be some advantages to this approach.

Now, I'm in the beginning process of rewritting an order tracking system that employed such a design and is again started to think about reasons why any developer would want to do this.

I know that there are numerous fields (name, address, phone, fax, email, etc.) that each entity shares in common that could be "centralized" but I always though working out the differences in each entity is not worth deploying such a design.

Thus, if anyone would like to share his/her own opinions and experiences for use of such a design, I would appreciate your comments.