Motivation for syntactic specifications UML metamodel says that a class has a name and attributes and methods. The notation description shows some examples on how to show these but gives only vague rules. It leaves too much freedom for tool implementors. The UML specification is messy. It gives 'basic notation' descriptions with 'presentation options for tools'. Are the given presentation options the only possible options of are they only examples? Example: a class box has at three compartments in order name, attributes, operations. A tool may extend the language by adding additional compartments (sic!). However, the ordering of compartments w.r.t tool specific compartments is not specified (examples show them after the basic compartments). Then, the mapping sections discuss the mapping of tool presentation options to the semantic model (example - ordering and suppression in list compartments)! Attributes may be shown using the syntax of a specific programming language. The mapping of such attribute declarations to the metamodel is not discussed. The notation guide gives style guidelines about typefaces and positioning of items in list and name compartmennts. These are only guidelines. However, sometimes typefaces are mentioned in the syntactic specification (example - in name compartment if the class is abstract the name appears in italics. This overrides any name font setting by a tool. It's strange for a language specification to talk about a basic language, tool specific extensions of the language, and the appearance of some language elements in an imaginary tool. There should be at least a reference tool implementation. It's a poor excusefor a language standard!