Re: [wss] editorial comments on WSS 1.0 core

[NISHIMURA Toshihiro <nishimura.toshi@jp.fujitsu.com>]

Thomas,

Thank you for your feedback.
In this case, it is not a normative part, so I'd like to pick up the
last one.

> * However, other mechanisms such as "principal name" are not necessarily
> unique and therefore such references might not be unique.


I have recalled another matter about normative statements: the use of
"RECOMMENDS".
In the core spec, "RECOMMENDS" is used in three sentences.
  L957
    this specification strongly RECOMMENDS ...
  L1218
    this specification strongly RECOMMENDS ...
  L1938-1939
    WSS: SOAP Message Security RECOMMENDS ...

RFC 2119 defines the adjective "RECOMMENDED" but not the verb
"RECOMMENDS". Shoud we rewrite these senteces?
---
NISHIMURA Toshihiro (FAMILY Given)
nishimura.toshi@jp.fujitsu.com
STRATEGY AND TECHNOLOGY DIV., SOFTWARE GROUP, FUJITSU LIMITED

At Mon, 6 Dec 2004 10:53:56 -0800,
DeMartini, Thomas wrote:
> 
> ] (10) Appendix B: SecurityTokenReference Model
> ] P53 L1813-L1814
> ] [original]
> ] | However, other mechanisms such as "principal
> ] | name" are not required to be unique and therefore such references
> may be
> ] unique.
> ] 
> ] I think the last part of this sentence should be "may not be unique".
> ] 
> ] [change to]
> ] | However, other mechanisms such as "principal
> ] | name" are not required to be unique and therefore such references
> may
> ] not be unique.
> 
> "may not" should be avoided when writing standards, because it is
> ambiguous whether it means 1) "you are not permitted to do X" (the
> actual English meaning) or 2) "you are permitted to not do X" (the
> meaning people often are trying to get at).
> 
> If we intend to say meaning #1, then we should say "MUST NOT".
> If we intend to say meaning #2, then we should say "NEED NOT".
> 
> As it turns out, "NEED NOT" and "MAY" are synonyms, so any of the
> following are normatively equivalent:
> 
> * However, other mechanisms such as "principal name" are not required to
> be unique and therefore such references MAY be unique.
> * However, other mechanisms such as "principal name" are not required to
> be unique and therefore such references NEED NOT be unique.
> * However, other mechanisms such as "principal name" are not necessarily
> unique and therefore such references MAY be unique.
> * However, other mechanisms such as "principal name" are not necessarily
> unique and therefore such references NEED NOT be unique.
> 
> Since they are all normatively equivalent, we can pick whichever one
> sounds best.  I prefer the latter one.
> 
> Now, if it is the case that our intent was not to make a normative
> statement, but only an informative one, then we should use the
> formulation "might not":
> 
> * However, other mechanisms such as "principal name" are not necessarily
> unique and therefore such references might not be unique.
> 
> &Thomas.
> 
> 
> To unsubscribe from this mailing list (and be removed from the roster of the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/wss/members/leave_workgroup.php.