After seeing how MoinMoin uses plain old text files to store their information, and seeing how the Python, Ubuntu, and FreeBSD wikis (all of which use MoinMoin) haven't been raped security-wise, why haven't file-based information systems displaced traditional SQL databases? The way I see it, software like PostgreSQL require significantly more bureaucratic procedure to modify information, and it doesn't seem like all of that is necessary. Sorry if I look like a fool for overlooking <obvious advantage>, but might someone be kind enough to explain one?
Databases are much faster for dealing with huge amounts of data, pretty much. They're designed for that. Though you could argue a filesystem in itself is a specialized database =P Also... what's a fiat file system?
Backup and restoration of a file-based system is considerably more difficult than a database-based system. Also, the hard disk only has to read in one file to access the entire database, meaning that data can be pulled and pushed more quickly than file based systems. And actually, file-based systems in stuff like MoinMoin are less secure than those using SQL databases as the backend. One potential security issue of file-based systems is that they entirely rely on web server or CGI interfaces to access data. What makes them seem more secure is the obscurity of web software implementations that use file-based information storage systems like MoinMoin. While the main reason IBM invented relational databases is due to file storage constraints, the other reason was to force a level of security into databases. Most relational databases tightly control what kind of data can be stored or retrieved. This level of security is independent of other layers or security, giving multiple levels of protection.
