Re: Publication request for draft-spinosa-urn-lex

[[e-mail] (Dale R. Worley)]

There are at least some technical questions that need to be addressed.

1) The ABNF in appendix A contains <national-name>, but it seems to be
intended to be what is described as <local-name> in section 4.3.

2) The use of % for escaping is incompatible with the generic syntax of
URNs:

      encoded   = "%" (1*6DIGIT / ("x" 1*6HEXDIG))  
                                 ; Hexadecimal digits (0-9, A-F)

as opposed to:

   pct-encoded   = "%" HEXDIG HEXDIG

3) The syntax for using % does not allow an encoded string to be parsed
unambiguously:

   jurisdiction-unit = alfanum *normal

   normal = alfanum / "."

      alfanum   = lowercase / DIGIT / encoded

      lowercase = %x61-7A        ; lower-case ASCII letters (a-z)
                                 ; DIGIT %x30-39 decimal digits (0-9)

      encoded   = "%" (1*6DIGIT / ("x" 1*6HEXDIG))  
                                 ; Hexadecimal digits (0-9, A-F)

For example, considering the string "x%1234" as a <jurisdiction-unit>,
it could be one <lowercase> followed by a <DIGIT> and three <encoded>s
("x", "%1", "2", "3", "4") or one <lowercase> followed by one <encoded>
("x", "%1234").

Perhaps this does not matter if it is not required that the underlying
UTF-8 string can be derived from the URN, but if so, then there is no
assurrance that two distinct UTF-8 string will be encoded as distinct
URN substrings.

Dale