GIS development essay (2) -- several questions about GIS database establishment

The real practice started. According to the user's needs, the first thing is to write a lot of such documentation instructions, and analyze and establish the database table and its content according to the user's needs.
During the database creation process, we used the Microsoft SQL Server2000 Database, and moved online some time before analyzing the database, I found that many people support Oracle databases (I have never used Oracle databases, but the leaders decided to use SQL Server2000 databases). I heard that the concept of spatial database support exists, you can put the whole map information into the database, but you don't know how much you know about it. At present, we use the SQL Server2000 Database to create tables related to the application system (such as permission control tables and user information tables) tables related to communication devices (for example, the instrument information table, which contains the upstream and downstream data or command formats of the instrument) and layer tables (for example: basic layer table and topic layer table. The tables in this section must correspond to layers one by one, and each table has a field that corresponds to the table file data in the corresponding layer one by one) and non-layer details tables (such: the managed information table and warning information table. It mainly refers to the Data Relationship between various parts of the system and the table used for system parameter settings). Do you know if this is the case? I also hope that people who know can give some answers.
There is also a question about the database. When I used SQL Server2000 to create an application database, I always wanted to add constraints between the table and the table, so that the data in the database can ensure data integrity, however, recently I was asked to remove these constraints from the physical sense, but they still exist in the logic sense. It is necessary to implement these logical constraints by programming and other methods. For example, a table is a department, a table is a person, a person is a department, and a department and a person are in a one-to-many relationship. In the past, I used to add a foreign key constraint between two tables: Personnel and department. Now I have to remove this foreign key, in addition to its primary key "Record ID", add a unique field "department ID" to the department, and add the same "department ID" field to the personnel table, the two tables do not have actual primary/foreign key constraints, but only have a logical relationship. I don't know which one is better than the previous practice. I personally think that although no actual constraint makes the table more flexible and more independent, it is better to add the primary and foreign key constraints. In most cases, users may want data accuracy, so it should be better to ensure database integrity. What do you think?
Recently, the database has been established. Now, we are using. Net to develop system function programs and public classes. After these tasks are completed, I believe we should focus on ArcIMS and mapnet.
However, the establishment of a spatial database is also the first time I have established such a database. If you have any incorrect opinions, please give your own opinions and experience.