Currently input and output IDs are effectively dumb string inputs. This means you can do things like put spaces in them etc. Since ids are such a common type of argument and they have inherent restrictions (can't duplicate, must be all alphanumeric, should trigger server updates if changed) then it makes sense we give ids a special class of argument. This will probably be reflected in a inputType: "ID" field for the argument info section of info json blobs.