I want to make the choice that skip reading the magic cookievalue and the schema description part and just assume I have one table with the specified fields and field lengths and that these will never change in name or size.
This way the reading part would be much easier but then I would lose the option to rename fields and sizes in the schema part. For those who is building up the databse dynamically according to the schema part the solution has more advantage but I want to keep the solution simple as possible,
do you lose points if you disregard the schema information?
I do not totally understand your question. You have to design and develop an application that handles the information from the db-file for the users. These information are only the records with content.
Your Implementation of the DataAccess of course needs to know the schema (field lengths,..) to select the required record for the user correctly.
I use for example a DbFileChecker at startup, to quarantine that the supplied db-File is a valid one. But in runtime I will always skip all the magicCode, number of fields, fieldName, length, because I would like to get the real stuff.
I think you mix two concepts here. 1.The magic Cookie - it helps you to identify the right database file 2.the meta data information / handling
You are free to hardcode (but I don't recommend this) or to build an dynamic model for the meta information (here I mean fields, fields size, field order, etc) as long as you justify your decisions. The mg. is used to identify the right database file if you skip it you may fail during reading(because the database follows an other scheme) you you can permanently damage an existent db (just by adding a record with a wrong format).