Naming convention should be less restrictive for enumeration literals

The current naming convention seems overly restrictive for enumeration literals. I n a modelling effort related to transport of dangerous goods, I encountered a lot of fixed lists that are actually sequences of numerical codes (e.g.: {"1", "2", "3", "4.1", "4.2}). The canonical encoding would be an enumeration with just these codes as literals. This is not allowed currently. I can't see why naming conventions for enumeration literals have to be that restrictive. In XML they are mapped to XML element content, which is fine. In other potnetial future platforms - that might not allow such character sequences as literals - a prefixing can be done generally in the mapping. Since the codes are actually domain concepts, it would be really preferable to have them as literals in the UML diagrams rather than having to introduce strange mappings at the data modelling level.