The following small example model illustrates the usage of parse rule fragments in Xtext (since 2.9). In earlier versions, often you had to repeat patterns — and make changes in all places where the same pattern was used.

Class: Modifiers "class" Named "{"

Field: Modifiers "field" Named Type;

Method: Modifiers "method" Named Type;

fragment Named: name=ID;

fragment Type: "type" type=ID;

fragment Modifiers: 

Note that the calls to Named etc. are done without assignment to a feature. If you look at the inferred classes Field etc. extend the interface Named.

Prior to 2.9 you may have written

Named: A|B;
A: name=ID;
B: name=ID;

in order to enforce a common supertype for A and B holding the name feature. However, if they share other compatible features, these may — unintentionally — end up in Named.