The idea is to have a PHP class for store information in "database tables" so using "records" and with the possibility to retrieve the information using "queries", without using any other server technology that the PHP functions for read/write files.
Advantatges:
the PHP file which contains de class definition and the text files which contains the "user-data" (tables, rows,...) can be in a unique directory. So the backup of the database is easy easy in a wide variety of ways! handly o automatizated.
the efficiency of the database is notably high and fast in some cases: not more than 5.000 records and in a shared hosting. Really is an optimal system to store/retrieve data for simple web applications.
the database only need the PHP file with the class (less than 30kb!!!) and the PHP file for "graphically" admin the database (like a "remake" of phpmyadmin ;) (less than 30kb also).
the "fields" of the tables don't need "type" definition like in tradicional databases (integer, date....). All the data is EVER stored as text. This make more easy and fast to program queries... although there exists then some limitations ;)
the content of each "field" in any "record" of any "table" can be so long you need!!!! you has not to define any more the length of your fields ;))
the queries are built using associatives arrays instead of the "limitant" classical SQL queries which uses text strings like: "where field1='value1'... "
Handicaps:
Not SQL queries :(( but let the user to specify the name of the table, "where" conditions (with some operators: =, <, >, LIKE, ...) joined with some "binding operators" (and, or).
not exists (at the moment) things like: indexes, vinculed fields, etc., typical of SQL databases.
Demo: