You want all "destructive" actions to be the result of a concious decision by the operator. You never want something "destructive" to happen by accident (as much as possible, anyway). So you make the ring a "write-enable" so that it requires a concious decision by some one that they want to destroy whatever contents is currently on the tape by adding the ring. If they forget to add the ring, the worst that'll happen is they'll get a "tape cannot be written to, it's not write-enabled" error message. If it was a "write-protect" ring, then worst that can happen is someone forgets to add the ring and overwrites critical data by accident.
It's a similar situation if you're programming a dialog box (say). Generally, you make the default button the "least destructive" action. For example, if you have a "Do you want to save before quitting?" dialog box, the options are "Save", "Don't Save" and "Cancel" - with "Cancel" the default. If someone just blindly hits "enter" then the program does the least destructive thing: doesn't save, and doesn't quit without saving.
Another analogy with web programming: you always put a "destructive" operation behind a POST (or DELETE or other non-GET) request. The idea is that if a bot comes along and crawls your site, you don't want it accidentally deleting posts or clearing data just by issuing GET requests: the destructive stuff is only initiated by a concious action by a human.