Re: [tag] Comments on the TA guidelines

[Stephen Green <stephen.green@documentengineeringservices.com>]

Responses below, inline

I agree to do at least one of these for next draft :-)

Best regards

Steve
---
Stephen D Green




2010/1/19 Jacques R. Durand <JDurand@us.fujitsu.com>:
> Comments on the TA guidelines
>
> -----------------
> [1]  I think we should make it more explicit that The "TA model" introduced
> in Section 3, is directly related  to the Test Assertion Model specification
> [TAM]. i.e. not just list [TAM] and [TAML] references in the ref section.
> This could be done beginning Section 3, saying something like:
>
> " The TA model described here is more formally defined in [TAM]."
>
> Such explicit refrence for the mark-up [TAML] has been made in 5.3 examples.

I agree


>
> -----------------
> [2] The example in 5.3 is badly formatted: each TA should be separated,
> maybe a better indent too.
>

There is sometimes a problem when wwe convert to (X)HTML
for the official committee draft and public review draft from ODT;
the conversion results in (X)HTML which exaggerates any
separations between lines. I will see what I can do but I had
bad experiences with this last time we went to public review
and I'm hoping removing all but essential newlines will help.
Indentations: again I'll see what I can do but formatting and
preserving it properly when converting between ODT and other
formats has proved to be a problem in the past.


> -----------------
> [3] The material in Appendix A should be bulleted or the like, to better
> separate each entry.

Again, conversion problems may limit what I can do, but I'll try

>
> -----------------
> [4] In the Conformance section:
> In addition to say that this is "non-normative" document, we should here
> again remind the reader that this guidelines doc is however about a
> normative specification ([TAM]), that must be complied with, when writing
> actual test assertions in a more formal way (i.e. based on a formal
> representation such as the XML representation in TAML).
> In addition, in case  the comment [20] I made for the "TA model" spec is
> agreed on (extending the notion of conformance to the model, to TA instances
> as well, even those using the "informal" notation), we should remind the
> users that even when using simply the informal notation used in this
> guideline documents, TA instances defined that way must still  conform to
> [TAM] (which should be the case if the reader follows this guidelines).
>

Maybe I'm misunderstanding but I think there are
many possible ways to write TAs (even usefully
applying our guidelines) which wouldn't even attempt
to conform to the TAG model so I'd not think we
should be implying anything negative about those.
There are formal ways to write test assertions to
which our guidelines apply which do not implement
our model or markup. I think we should leave this as
it is and ensure we say nothing about conformance
(we don't even know whether the model/markup will
be accepted as a standard and even then the
guidelines would still be useful as an informative
document).


>
> -jacques