[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [tag] Re: Proposed shorter text for 3.7.2 Composition of Assertions
Thanks Kevin I'll probably go back to the original and start from there with your proposals; then'll see if anything else seems appropriate to change too, in line with those changes. Best regards Steve 2009/1/8 Kevin Looney <Kevin.T.Looney@sun.com>: > Hi Stephen, > > Here are some comments from your last edit (and merging some of the previous > comments from Jacques/me). > > Thanks, > Kevin L > > ________________________________ > > ([KL]Stephen, we should probably look back and edit the intro to 3.7. If we > remove the section on visibility, then we can probably remove the breakdown > description of the following subsections. ) > > 3.7.2 Composition of Assertions > > There are three dimensions that describe how assertions from a > referenced specification may be included within an umbrella > specification: > > 'scope of inclusions', > 'conditionality of inclusions', > 'modification of inclusions', > > > These relationships between specifications can be expressed by > including a set of references in the test assertions document for > the umbrella spec. The references would be to any test assertions > from the referenced specification which are required according to > the conformance clause of the umbrella specification. An alternative, > if no such test assertions for the referenced specification exist, would > be to summarize conformance requirements to the referenced > specification in the form of a test assertion using its predicate: > > ([KL]Jacques revised the example - introducing normative statement, and > identified requirement.) > > Case 1: The Widget specification says: "All requirements in this section > only apply to "mini" widgets, i.e. widgets > that are conforming to the Mini-Widget Small Box specification 1.2." > > "Requirement 701: If a mini-widget has a battery holder, then the > mini-widget MUST be labelled as 'low voltage'." > > > TA id: widget-TA701 > Target: Widget > Normative Source: Requirement 701 of the WidgetSpec 1.0 > Predicate: [the widget] is labelled 'low voltage'. > Prescription Level: mandatory > Prerequisite: [the widget] is conformant to the Mini-Widget Small Box > Specification 1.2 AND [the widget] has a battery holder. > > Scope of Inclusions > > An umbrella specification usually relates to a referenced > specification by assuming or requiring conformance of its > implementation to this specification. [KL]These conformance relationships > may be expressed in either the predicate or the prerequisite of a TA. > Variations exist as follows. > > ([KL]reduced to two variations.) > > conformance to an (entire) umbrella specification, or specification profile > conformance to a specific normative statement from the umbrella > specification > > The predicate or the list of external test assertion references > would reflect such variations. > > ([KL]I think this statement is meant to replace the descriptions of how one > might express conformance (in either predicate or prerequisite) to either > (a) whole/part of a referenced spec, (b)single normative statements in a > referenced spec. The section is definitely 'more terse' without the > enumeration of these variations. Is it 'too far' without the examples? > Perhaps TA701 shows a proper example (eg. Conformance expressed in the > prereq to a (entire) referenced spec)?) > > Conditionality of Inclusions > > It might be that the conformance or otherwise to a referenced > specification is a condition which is included in another test > assertion as a prerequisite. The prerequisite of the assertion may: > a. require that optional portions of the referenced specification > be implemented in the umbrella, > b. conditionally require optional portions of the referenced > specification be implemented in the umbrella (for example, based on > the presence of hardware or some other such support), or > c. make remaining (required) portions of the referenced > specification optional. > > ([KL]Jacques and I considered rewording this (perhaps orthogonally). I > think more emphasis should be placed on how 'referenced specs' have profiles > with various statements of inclusion/optionality. The conditionality applies > to 'overriding the optionality of the profile'. > > The key is that 'when an umbrella spec references these profiles, it may > specify a change to this inclusion/optionality in a prerequisite in various > ways (mandatory -> optional, optional -> mandatory, optional/mandatory > (based on other conditions)). > > ([KL] I also wonder whether this section is too terse without an example.) > > Modification of Inclusions > > This dimension of inclusion describes where an umbrella specification > is conformant to a referenced specification, where some subset of > assertions must be modified. This means of inclusion assumes some > partitioning of the unchanged assertions and modified assertions. You > can use "lists of assertions" to describe in the prerequisite the > subset of assertions that the umbrella specification is conformant to > "unchanged". The remaining test assertions (the changed set) can be > individually specified as test assertions of the umbrella > specification. > Typically, assertions are modified in a referenced specification that > can be strengthened in a few ways: > > strengthening the prescription level of an assertion (eg. x MAY do y > => x MUST do y), or > strengthening the meaning of an assertion with additional > requirements (eg. IF x THEN z => IF (x AND y) THEN z). > > ([KL] This section remains at a sufficient level of description.) > > > ([KL] Jacques has claimed this new dimension, we haven't really discussed it > much. > > Jacques description:) > > For example, Spec B is a > specialized form of Spec A. Here, spec A is most general spec, and spec B is > adding to it (e.g. like a WS-* spec is adding to or building on top of SOAP > spec). So here, B is often requiring full conformance > to A. > > and it looks like Stephen introduced this in 3.7 intro: > > Specification modularity may also come from a "prototypical specification" - > a "base specification" - from which specialized derivative specifications > may define specific information for a given context. > > > ([KL] Upon further thought, I'm wondering if this is a dimension, if it is a > sub-case of "Modification", or maybe just something different. > > Consider this as a sub-case of Modification: With the addition of more > assertions to a specification, we are typically 'strengthening' the > specification (in a similar way to modifying individual assertions to > strengthen requirements). > > Jacques, if you agree with this view, perhaps we can simply mention in the > modification section that 'a referenced specification may be modified > through the addition of further requirements'.) > > > > > > > ________________________________ > > > Stephen Green wrote: > > " 3.7.2 Composition of Assertions > > There are three dimensions that describe how assertions from a > referenced specification may be included within an umbrella > specification: > > 'scope of inclusions', > 'conditionality of inclusions', > 'modification of inclusions'. > > These relationships between specifications can be expressed by > including a set of references in the test assertions document for > the umbrella spec. The references would be to any test assertions > from the referenced specification which are required according to > the conformance clause of the umbrella specification. An alternative, > if no such test assertions for the referenced specification exist, would > be to summarize conformance requirements to the referenced > specification in the form of a test assertion using its predicate: > > TA id: widget-TA108-1 > Normative Source: [interpretation of conformance clause to WidgetSpec > 1.0] "All widgets conformant to the WidgetSpec 1.0 specification must > also be conformant to the 'smaller box' assertions of the WidgetMobile > Small Box Specification 1.2" > Target: widget > Predicate: [widget] Conforms to WidgetMobile Small Box Specification 1.2 > Prescription Level: mandatory > > > Scope of Inclusions > > An umbrella specification usually relates to a referenced > specification by assuming or requiring conformance of its > implementation to this specification. Variations exist as > follows. > > conformance to an (entire) umbrella specification > conformance to a profile of the umbrella specification > conformance to a specific normative statement from the umbrella > specification > > The prediacte or the list of external test assertion references > would reflect such variations." > > > Conditionality of Inclusions > > It might be that the conformance or otherwise to a referenced > specification is a condition which is included in another test > assertion as a prerequisite. The prerequisite of the assertion may: > a. require that optional portions of the referenced specification > be implemented in the umbrella, > b. conditionally require optional portions of the referenced > specification be implemented in the umbrella (for example, based on > the presence of hardware or some other such support), or > c. make remaining (required) portions of the referenced > specification optional. > > > Modification of Inclusions > > This dimension of inclusion describes where an umbrella specification > is conformant to a referenced specification, where some subset of > assertions must be modified. This means of inclusion assumes some > partitioning of the unchanged assertions and modified assertions. You > can use "lists of assertions" to describe in the prerequisite the > subset of assertions that the umbrella specification is conformant to > "unchanged". The remaining test assertions (the changed set) can be > individually specified as test assertions of the umbrella > specification. > Typically, assertions are modified in a referenced specification that > can be strengthened in a few ways: > > strengthening the prescription level of an assertion (eg. x MAY do y > => x MUST do y), or > strengthening the meaning of an assertion with additional > requirements (eg. IF x THEN z => IF (x AND y) THEN z)." > > > > > > > > > -- Stephen D. Green Document Engineering Services Ltd http://www.biblegateway.com/passage/?search=matthew+22:37 .. and voice
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]