Re: Testable assertion tagging for W3C specifications

On Thu, 30 May 2002, David Marston/Cambridge/IBM wrote:

> Based on Alex Rousskov's latest description, I surmise that he is
> proposing:
> 
> 1. Spec authors take no special action.

Spec authors MAY use cool markup provided by QAWG.

> 2. Test case authors cite the specs using whatever crude means are
>    available. They may have to use byte counts to point at particular
>    characters.

Test case authors MAY use cool addressing schemes provided by QAWG to
cite the specs. If QAWG does not provide applicable cool addressing
schemes, or the authors neglect to use them, the authors have to use
whatever means are available, including crude ones, naturally.

> 3. Somebody applies a process which takes as input the specs and the
>    test cases (or whatever is doing the citing) and produces new
>    versions of the specs with extra footnote marks.

Yes, if the intent is to link the spec with test cases.

>  Each test case puts its own footnote marks.

If the intent is to link the spec with individual test cases, then
each individual test case is linked using a single footnote mark. If
the intent is to link the spec with groups of test cases, then one
group is linked using a single footnote mark.

There may be many renditions of the spec, depending on the intent, of
course.

> 4. The reader must read the specially-processed version of the spec.

The reader must read the specially-processed version of the spec if
the reader wants to have access to special features offered by that
version, naturally.

>    upon encountering a footnote mark, s/he may click on it and see the
>    test case that supplied that mark.

Yes, where "click on it" may mean other actions, depending on the
interface; all resulting in getting details about information being
linked from the specs (that information can be a single test case
description, a description of a group of test cases, a test suite
description, etc., etc.). It is just a link.

> The description of this capability:
> > >  - we want to provide reverse index where a reader can
> > >    read the specs and get "one-click" access to
> > >    test cases that correspond to the sentences she is reading
> I took that to mean that one click would fetch all the tests, but I
> now think it means that you may have several marks side-by-side, and
> you must click on each mark to see one test. So the special edition
> of the spec might look like this:
> 
> If the attribute name matches PrefixedAttName, then the NCName gives
> the namespace prefix, used to associate element and attribute names
> with the namespace name in the attribute value in the scope of the
> element to which the declaration is attached. @@@@@@@@@@@@In such
> declarations, the namespace name may not be empty.
> 
> Each "@" would link to a separate test case.
> 
> Alex: Is this what you meant?

Yes and no. You are citing one possible application of the technology.
Whether one mark corresponds to one test case, or a group of test
cases, or something very different is not specified by the addressing
scheme and does depend on the intent/interface.

Alex.

Received on Thursday, 30 May 2002 14:28:48 UTC