RE: [tag] Draft write-up for Anatomy of a TA section

["Durand, Jacques R." <JDurand@us.fujitsu.com>]

Kevin:

It would be good to complement this draft (written in a tutorial mode),
with a real example of pre-requisite from Java cases (if it is possible
to isolate one as a small standalone unit, easy to understand for most
readers).
 
Other special cases of pre-requisites (like Lee's "implicit
pre-conditions", or global pre-requisites) might need to be addressed
differently, we don't know yet. 

A possible structure for the TAGdoc is that it could be centered on a
short "tutorial-like" general intro structured around a running example,
and that skims through the main notions (parts of a TA, granularity of a
TA, pre-requisites, test environment,...) while establishing basic
methodology. Then this tutorial can refer to more specialized sections
that develop on each concept, and illustrate them further.

Jacques

-----Original Message-----
From: Kevin.T.Looney@Sun.COM [mailto:Kevin.T.Looney@Sun.COM] 
Sent: Tuesday, August 28, 2007 8:47 AM
To: Durand, Jacques R.
Cc: tag@lists.oasis-open.org
Subject: Re: [tag] Draft write-up for Anatomy of a TA section

Hi Jacques (et. al).

     I just wanted to quickly comment on something I liked.  On page 3,
while discussing example TA-102, there is a notion of  "pre-requisite" 
test assertions.

     Capturing the assertion pre-requisites (and thus, dependency
relationships) looks like it could be very important and practical.  For
Java conformance testing, we have large (10-40k test) test suites. 
 Running these can take significant time.  Pruning unnecessary tests (to
reduce testing cost) becomes an important usability factor for these 
test-suites.   A testing harness that can delete trees of unnecessary 
tests (due to pre-requisites failing) would be very useful in this
scenario.

      From the perspective of specification analysis, identifying a Test
Assertion (by distilling normative text into it's most atomic behavioral
description) can be problematic if "the context" of the TA is removed. 
 Specifying the pre-requisite assertion is a step in the right direction
here.

Thanks,
Kevin L

Durand, Jacques R. wrote:

> Authored by Serm and myself, although Serm has not seen this latest 
> version as he is traveling right now.
> For comments.
>  
> -Jacques