The native form for RFA identifiers is info:rfa/*. There are 
no legacy forms. 

Identifiers in the rfa namespace are used to identify 
'items' in 'collections'. The syntax of the identifier is of 
the form:

  info:rfa/<registryLabel>/<collectionLabel>/<itemLabel>

where:

  * <registryLabel> is a label for a registry that can be 
used to facilitate interoperability across BaseURLs. This 
label MUST be registered in the service-mechanism. Although 
these labels must be registered, it is NOT required that 
they be connected to a resolver. Note! The labels 
"webservices" and "localhost" are reserved for internal 
purposes.

  * <collectionLabel> is a label for a collection within a 
registry. Note! The label 'collections' is reserved for 
internal purposes.

  * <itemLabel> is a label for an item (unique within a 
collection).

Labels in an RFA identifier may contain the following 
characters:

()*+,-./;=?@
0-9
A-Z
a-z

i.e. 12 punctuation characters, 10 digits, 26+26 upper and 
lowercase English letters. An RFA identifier does not 
contain whitespace characters.

The only character in an RFA identifier that needs to be 
%-encoded within an "info" URI is the "?" character which 
should be %-encoded as "%3F", i.e. all instances of "?" 
should be replaced by the string "%3F" 

Note! Only itemLabel is  allowed to contain the '/' 
character. Otherwise the URI pattern couldn't be parsed 
unambiguously.

RFAs are case-insensitive.

An RFA (Registry Framework Architecture Identifier) is a 
string of characters capable of uniquely representing an 
item in a collection. The identifier is a string of unique 
characters of indefinite length according to the needs of 
the owner of the <registryLabel>. It can be used for both 
public and private databases. It is extensible and is a 
non-proprietray identifier which can be used in both printed 
and electronic databases.

RFAs can generally be resolved via the service-mechanism by 
replacing "info:rfa/" with "http://purl.org/rfa/".