... and if, by chance, you're seeking to clean up user input from the web, please, don't stop with character entities. Read up on untainting for a start, if for no other purpose than to get some flavor of the scope of security issues with un-moderated user input.

Cleaning up html entities -- even as benign as   or &sect (though, in some instances, those could be abused) -- is often a good thing... but depending on the source of the "data" you're seeking "to sanitize," it may be only an inadequate protection.