One reason to use the content repository is that it already has the logic necessary to differentiate between Oracle CLOB and Postgres TEXT storage. It seems silly for individual packages to duplicate such functionality.
Another reason to use it is that we plan to extensively improve the Tcl API to the CR in the upcoming months, making it easier to use.
Of course others have mentioned other valid reasons for using the CR.
Versioning is useful in many, many more contexts than in a CMS, BTW. Surely the capability is not always needed and I've looked into the possibility in the future of allowing non-versionable content types that use a single object for both the cr_item and cr_revision as a way of lowering overhead in such cases. But the CR brings more to the table than just versioning.
You're free to ignore it in your own work, of course.
