NEW ISSUE: 2119 terms apply to implementations,not message (document) instances

[Doug Bunting <Doug.Bunting@Sun.COM>]

*title*: 2119 terms apply to implementations, not message (document) 
instances

*description*: RFC 2119[1] assigns very specific meanings to the words 
"MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD 
NOT", "RECOMMENDED", "MAY", and "OPTIONAL".  We say this RFC applies to 
our specification but do not consistently use the words as defined.

RFC 2119 is about requirements of a specification (for implementations 
of said specification that is) and not about cardinality in or other 
constraints upon an XML message (or document in general) instance.  
Phrases in the RFC such as "particular behaviour is acceptable" and 
"implementation which does not include" make this distinction quite 
clear.  We should use other terms to describe cardinality &c constraints 
for the elements and attributes in our schema.

*justification*: Do not want to confuse readers using 2119 terms in ways 
which are distinct from their RFC 2119 definitions.  In other words, to 
avoid complaints from the spec weenies :)  May have the general benefit 
of improving readability.

*target*: core (both specifications)

*type*: editorial, perhaps after some discussion in TC giving the 
editorial team some direction

*proposal*: Editors run through the specifications and correct phrases 
such as

    * "the action IRI MUST consist of the WS-RM namespace" (line 122)
    * "Additional children elements ... MUST NOT contradict the
      semantics" (lines 229-230)
    * "This element MUST NOT be sent as a header block" (line 255)
    * "This OPTIONAL element" (line 545)

and correct them to describe the constraints in terms of the 
implementation.  In many cases, this change will have the additional 
benefit of changing the voice from passive to active.

*related issues*: i040, "Change 'optional' and 'required' in section 3 
to RFC 2119 OPTIONAL and REQUIRED" which exposed additional RFC 2119 
terminology "violations"

*referenced specification*: WS-RM CD II[2]

[1] <http://www.faqs.org/rfcs/rfc2119.html>
[2] 
<http://www.oasis-open.org/apps/org/workgroup/ws-rx/download.php/16271/wsrm-1.1-spec-cd-02.pdf>