Validate schema EBML IDs

[StokesMIDE]

When importing a schema, it would be fairly simple to verify that the IDs are the correct length for the number of preceding zeros in their first byte. This would help developers working on new schemata. A badly structured ID could raise a meaningful exception, like: ValueError: Bad ID (0x2ABC): A class C EBML ID must be 3 bytes.

Arguably related to #32.

[StokesMIDE]

Suggestion: maybe some sample code somewhere to generate IDs? Maybe in util.py. A function kind of like this:

def generateID(class_='a', schema=None, exclude=[]):

The optional schema is used to generate an unused ID. Similarly, exclude is a list of IDs, so the function can be used multiple times without updating the schema.