[1] [[URN]] 亜空間 [CODE(URI)[[[urn:x-suika-fam-cx]]:msgpm]]
は、 [[manakai]] が内部で使用する名前のためのものですが、
もちろんそれ以外でも使用することができます。
主として [[822]] 的メッセージの構造の部分を識別するのに使うことができます。

[2]
,urn:x-suika-fam-cx:msgpm:header:default:[VAR[field-name]],頭欄 (既定名前空間)
,urn:x-suika-fam-cx:msgpm:header:http:[VAR[field-name]],[[HTTP]] [[頭欄]]
,urn:x-suika-fam-cx:msgpm:header:http:c:[VAR[field-name]],HTTP 頭欄 ([CODE(HTTP)[C-[VAR[*]]:]] 欄),Deprecated [2005-05-19]
,urn:x-suika-fam-cx:msgpm:header:http:cgi:x:[VAR[field-name]],[[CGI頭欄]] ([CODE(CGI)[X-CGI-[VAR[*]]:]] 欄),Deprecated [2005-05-19]
,urn:x-suika-fam-cx:msgpm:header:mail:rfc822:[VAR[field-name]],[[RFC 822]] 頭欄
,urn:x-suika-fam-cx:msgpm:header:mail:rfc822:content:[VAR[field-name]],[[MIME]] 欄,Deprecated [2005-05-19]
,urn:x-suika-fam-cx:msgpm:header:mail:rfc822:list:[VAR[field-name]],RFC 822 [CODE(822)[[[List-*]]:]] 欄,Deprecated [2005-05-19]
,urn:x-suika-fam-cx:msgpm:header:mail:rfc822:original:[VAR[field-name]],RFC 822 [CODE(822)[[[Original-*]]:]] 欄,Deprecated [2005-05-19]
,urn:x-suika-fam-cx:msgpm:header:mail:rfc822:resent:[VAR[field-name]],RFC 822 [CODE(822)[[[Resent-*]]:]] 欄,Deprecated [2005-05-19]
,urn:x-suika-fam-cx:msgpm:header:message:delivery-status:[VAR[field-name]],[[DSN]] 欄
,urn:x-suika-fam-cx:msgpm:header:message:disposition-notification,[VAR[field-name]]:[[MDN]] 欄
,[CODE(URI)[urn:x-suika-fam-cx:msgpm:header:mime:]],[[MIME頭欄]],Assigned [2005-05-19]

[3] [CODE(URI)[[VAR[field-name]]]]s are case-insensitive. 
Lowercase is the canonical form and is recommended.
[SRC[Resolution 2005-05-19]]

[4] Some header field might be encoded in multiple forms of URIs.

For example, an [[RFC 2822]] header field [SAMP(822)[List-ID]]
is representable as both <urn:x-suika-fam-cx:msgpm:header:mail:rfc822:list-id>
and <urn:x-suika-fam-cx:msgpm:header:mail:list:id>.
In this case, the [CODE(URI)[urn:x-suika-fam-cx:msgpm:header:mail:rfc822:list]]
is deprecated and the later URI should be used.

Another example.  An RFC 2822 message header might have a header field
of name [SAMP(822)[[[Content-Language]]]].   URI reference of it might
be <urn:x-suika-fam-cx:msgpm:header:mail:rfc822:content-language> or
<urn:x-suika-fam-cx:msgpm:header:mime:content-language>.
Both URIs are valid and you can use which is appropriate (or useful).

[SRC[Resolution 2005-05-19]]

[6] '''Use of percent-encoding''':
As in [[RFC 3986]], any [CODE(ABNF)[[[unreserved]]]] octets may  
safely be percent-encoded and decoded, while the decoded form
(i.e. using the [CODE(ABNF)[[[unreserved]]]] character itself is recommended).
If a [CODE(URI)[[VAR[field-name]]]] contains non-[CODE(ABNF)[[[unreserved]]]]
character, then the character must be percent-encoded.
[SRC[Resolution 2005-05-19]]

[5] '''Using these URI references as XML vocabulary'''.
These URI references might be encoded as [[QName]]s and
be part of XML document type.  On such an information item, 
the URI reference minus [CODE(URI)[[VAR[field-name]]]]
is [CODE(InfoProp)[namespace name]] and the [CODE(URI)[[VAR[field-name]]]],
with all latin letters in lowercase, is [CODE(InfoProp)[local name]].
See [[全文検索用中間形式]>>6] for example.

NOTE: If a [CODE(URI)[[VAR[field-name]]]] has one or more characters
that is inappropriate as in an XML [[Name]], then the URI
reference is impossible to represent in the XML form.

[SRC[Resolution 2005-05-19]]

[7] '''Equivalence''':
In addition to the equivalence algorithms defined for 
[CODE(URI)[[[urn:x-suika-fam-cx:]]]] URI references,
there is an equivalence level 
<http://suika.fam.cx/~wakaba/archive/2005/5/equiv-header-field-name>
[SRC[Resolution 2005-05-19]].

- If two URI references start with
[CODE(URI)[urn:x-suika-fam-cx:msgpm:header:]]
(for [CODE(URI)[urn]] and [CODE(URI)[x-suika-fam-cx]]
any case-combinations are allowed), 
-= Converts all capital Basic Latin letters
([CODE(char)[[[LATIN CAPITAL LETTER A]]]] to 
[CODE(char)[[[LATIN CAPITAL LETTER Z]]]]) to
the lowercase.
-= Compares two URI references in character-by-character basis.
- If one of URI references does not start with
[CODE(URI)[urn:x-suika-fam-cx:msgpm:header:]],
this algorithm does not define the equivalence.
-- NOTE: Any other algorithm combined with this algorithm
may or may not define such equivalence.
- Note that [CODE(ABNF)[[[unreserved]]]] characters are always
percent-encodable.  It might be required to decode
before the test.