A PRONOM Unique Identifier (PUID) is composed of two 
elements, the PUID type, and the actual identifier. The PUID 
type identifies the class of representation information to 
which the identifier refers, where each identifier is unique 
within that class. Thus, both elements are required to 
construct a unique identifier.

The syntax may be expressed in formal BNF notation as 
follows:

<puid> ::= <puid_type> '/' <identifier>

<puid_type> ::= 'fmt'

<identifier> ::= <fragment> | <identifier> <fragment>

<fragment> ::= <digit> | <letter> 


<letter> ::= 'a' – 'z'

<digit> ::= '0' - '9'

The PUID type is expressed as a lowercase alphanumeric 
string. The only PUID type currently defined is "fmt", which 
denotes that the identifier relates to a file format. Other 
types may be defined in the future. The identifier itself is 
expressed as a lowercase alphanumeric string of arbitrary 
length. The PUID type and identifier are separated by a 
forward slash.

The PUID type is case sensitive and must be normalised to lower case.

The PUID namespace consists of identifiers which correspond to
entries in the PRONOM technical registry. Every identifier is unique within
the namespace, and is persistently and unambiguously bound to a single registry
entry. The PUID type element identifies the type of registry entry (e.g. a file
format), and the identifier element denotes an instance of that type within the
registry (e.g. the record for a specific file format).