RE: [AM] Why Traceability?

(responding to John)

> I was at a talk a few months ago where Ken Schwaber gave
> a quick introduction to Scrum and its place in the Agile
> universe. I remember him mentioning that Scrum was CMM
> level 2 and 3 compliant but now that I know a little more
> about CMM, I'm having difficultly understanding how an Agile
> process can be CMM compliant without losing its agility.
>
> My difficulty arises from CMM's "Maintain Bidirectional
> Traceability of Requirements" specific practice. As far as I
> can tell, you can't achieve universal traceability without
> heavy documentation or tools, both of which are discouraged
> in Agile processes. Automated tests (JUnit, NUnit, etc.)
> certainly help agility and give great traceability but they aren't
> appropriate (AFAIK) for testing user interfaces so user
> interface requirements must be traced some other way,
> which brings us back to documentation or tools.
>
> How do people satisfy CMM's need for Bidirectional
> Traceability without being crushed by documents or 
> distracted by tools?

First, let me correct a possible misconception.  Agile 
approaches are not absolutely anti-tool.  The agilists are most
enthusiastic about use of appropriate tools, and are pretty
prolific at the creation of appropriate tools (as all Unix
programmers used to be, way back... but don't let me get
distracted!).  Agilists don't like big, unwieldy, expensive
tools that require big, unwieldy, expensive process to go
along with.

The next question one needs to ask is, "What does CMM
mean when it talks about Bidirectional Traceability?"
Generally, the CMM auditors will take any reasonable
interpretation, as indicated in the earlier response on this
thread from Hubert.  If the test *is* the requirement, then
a passed test indicates the requirement traces to the code.
A reasonable code coverage tool will trace the code to
the requirement (in desperation, one could comment out
the code in question and see what tests fail).

Other tools can be used for testing the user interface - there's
a yahoo forum specifically on the topic of doing test-first
development for user interfaces, if you need help...
http://groups.yahoo.com/group/TestFirstUserInterfaces

So to answer the question in your final paragraph - you
use tools that don't distract.


Paul Oldfield

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
www.aptprocess.com

any opinions expressed herein are not necessarily those of
Mentors of Cally or the Appropriate Process Movement
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

For more information about AM, visit the Agile Modeling Home Page at www.agilemodeling.com
--^----------------------------------------------------------------
This email was sent to: [email protected]

EASY UNSUBSCRIBE click here: http://topica.com/u/?bUrKDA.bWnbtk.Z2NtYS1h
Or send an email to: [email protected]

For Topica's complete suite of email marketing solutions visit:
http://www.topica.com/?p=TEXFOOTER
--^----------------------------------------------------------------