I strongly believe this is the right approach! Anyone could choose any "delimiter", but the HTTP character-encoding approach is the right one, because it's already agreed-upon and to retrieve the '/' back requires nothing more than an already-existing
http_entities_decode() call!