Support for types

[dwalluck]

Right now, type validation/normalization is just a big method with a switch statement.

Should type handling be moved to separate classes, e.g., have PackageURL handle validation/normalization only in the general case and have each type handle itself?

[jeremylong]

Yes, this should be moved to a new class. This will be especially true once the spec is updated to have the types in a machine-readable format. Moving this to it's own class now will be a good step in preparation for the future.

[ppkarwasz]

Do we know how the spec will handle new type registrations? If the enumeration of types will not be centralized, a SPI might be useful to register new types.