Barry Wise on five progressive steps to database normalization
We start out with an example where we store user's name, company, company address, and personal  urls - say url1 and url2
The zero form is when all of this is in a single table and no normalization has occured.
The first normal form is achieved by
1) eliminating repeating groups in individual tables
2) creating a separate table for each set of data
3) identify each set of related data with a primary key
so this yields a table where the user information is repeated for each url so url field limitation is solved
The Second normal form is achieved by
1) Creating separate tables for a set of values that apply to multiple records
2) Relate these tables with a foreign key
Basically, we break the url values into a separate table so we can add more in the future
The third normal form is achieved by
1) eliminating fields that do not depend on the key
Company name and address have nothing to do with the user id, so they are broken off into their own table
The fourth and higher form depend on data relationships involving one-one, one-to-many and many-to-many.
The Fourth normal form is
1) In many to many relationship, independent entities cannot be stored in the same table.
To get many users related to many urls, we define a url_relations where they user id and url id are paired.
The next normal form is the Fifth normal form which suggests that
1) The original table must be reconstructed from the tables into which it has been broken down. This is a way to check that no new columns have been added.
As always, remember that denormalization has its benefits as well.
Also, Litt's tips additionally mentions the following :
1) create a table for each list. More than likely every list will have additional information
2) create non-meaningful identifiers.
This is to make sure that business rule changes do not affect the primary identifier