opencsa-liaison — archive

Use of "may", "must", etc.

From: Anish Karmarkar <[email protected]>
To: [email protected]
Date: Mon, 10 Dec 2007 08:50:00 -0000

> I think it is perfectly possible to write meaningful text without using > RFC 2119 terms. Whether we want to spend the time doing this > wordsmithing across > the specs is the issue. I agree. I also want to comment on the use of 'MAY' in the 1st example. I think it is appropriate. On other SCA calls, comments have been made that 'MAY' is all about interop. If you look at RFC 2119, it may^h^h^hmight seem that way, but in most practical implementation cases, the interop part results in impl. just raising errors when they don't support the optional part. The more important part (to me) about 'MAY' is that: it is optional, you don't *have* to implement it, but if you do then you MUST follow all the rules that are defined. -Anish -- Mike Edwards wrote: > > Folks, > > OK, I'll bite on the challenge set by Bryan for these two cases, in the > full and certain knowledge that showing a solution for > just these two will not address the wider sea of cases spread across all > the SCA specifications. > > 1) > I'd take this as being normative in the first instance. It is stating > something testable - that a C++ service interface has an optional > annotation, that, when present, indicates that the contract defined by > the interface is conversational. The target for the normative > statement is the interface document itself: > > "Service interfaces MAY be annotated to specify whether their contract > is conversational, as described in the Assembly Specification [ASSEMBLY], > using the @Conversational annotation." > > ...if you don't want the statement to be normative, then I would cast it > as follows: > > "A @Conversational annotation is optionally applied to a C++ Service > interface to indicate that the service contract is conversational, as > described in the SCA Assembly specification [ASSEMBLY]." > > 2) > I'm assuming that there is nothing normative here. > > "The data exchange semantics for calls to local services is > /by-reference/. This means that code needs to be written with the > knowledge that changes > made to parameters (other than simple types) by either the client or the > provider of the service can be seen by the other." > > > I think it is perfectly possible to write meaningful text without using > RFC 2119 terms. Whether we want to spend the time doing this > wordsmithing across > the specs is the issue. > > Yours, Mike. > > Strategist - Emerging Technologies, SCA & SDO. > Co Chair OASIS SCA Assembly TC. > IBM Hursley Park, Mail Point 146, Winchester, SO21 2JN, Great Britain. > Phone & FAX: +44-1962-818014 Mobile: +44-7802-467431 > Email: [email protected] > > > *Bryan Aupperle