Necessity of using rdfs:Datatype in RDF code

In a text book (A Developer’s Guide to the Semantic Web by Yu, Liyang), in page 143, it suggests that, "it is always useful to clearly declare that a given URI represents a datatype, as shown here:". Then it shows the following code:

<rdf:Property
    rdf:about1⁄4"http://www.liyangyu.com/camera#model">
<rdfs:domain rdf:resource="#Camera"/>
<rdfs:range
    rdf:resource="http://www.w3.org/2001/XMLSchema#string"/>
</rdf:Property>
<rdfs:Datatype
    rdf:about="http://www.w3.org/2001/XMLSchema#string"/>

I fail to understand the necessity of the last two lines. As far as my knowledge goes, XMLSchema#string will be good for any string. And as is delineated here, rdfs:Datatype is "the class of datatypes".

What do we lose if we do not use the last two lines?

Solution

All primitive types differ, though reasoners may or might think in the opposite way.

The supertype (⊆) of all the primitive types is rdfs:Literal, which is not to be confused with rdf:PlainLiteral, and which is analogous to xsd:anySimpleType.

The type (∈) of all the primitive types is rdfs:Datatype.

In the previous edition (2011), Liyang Yu wrote right after the piece you quote:

The next example will show that using rdfs:Datatype is not only a good practice, but also necessary in some cases.

I believe, this example wasn't excluded in 2014 edition.