Hello,

I'm trying to design a database structure for the purpose of keeping track of recent programs and their features.

* Each program is organized in a category and section.
* There is a set of pre-determined features that any program may have.

Example of features:

Language Written in:
C/C++
Assembly
Visual Basic
Java

General Features:
Pluggable
Source code
Documentation
Price

Platform:
Windows
Linux
MacOS

etc....

When listing the programs, a tick beside the feature would indicate that a featuere is supported.

I would like to design the database such that any feature can be added/removed easily at any time (like for example adding Solaris to the list above for features supported).

I initially thought of having the following tables:

ProgramEntry
id, author, name, description, addDate

FeaturesData
prog_id, feature1, feature2, feature3, ....

And whenever I want to add support for a feature in general, I would just alter the table to add a column.

However, this design is not that scalable and flexible. For example, if I would like to also store the date of when a feature of a program was added, then that's not possible with the table design I provided.

Does anyone have any suggestions of how I would design the tables such that it incorporates the ideas I just mentioned?

Thanks for everybody's help.

I would recommend a more flexible structure. I would design it this way:

relations:

program m:n features
program m:n platform
program m:n author

so get seven tables:

table -> columns -> Primary key -> foreign key
program -> id, name, description -> id
features -> f_id, feature_name -> f_id
platform -> p_id, platform_name ->p_id
author -> a_id, authors_name -> a_id
program_author -> id, a_id -> (id,a_id)->id, a_id
program_features -> id, f_id -> (id,f_id) -> id, f_id
program_platforms -> id,p_id -> (id, p_id) -> id, p_id