[cgl_discussion] Words in requirements

Pradeep Kathail pkathail at cisco.com
Tue Oct 1 22:03:11 PDT 2002

I will suggest using:
"OSDL CGL specifies that carrier grade Linux shall provide..."


At 9/30/2002 05:34 PM -0700, Peter Badovinatz wrote:
>Issue, as described in the minutes to today's requirements call:
>1)Each requirement says "OSDL CGL shall provide [or develop, etc]..."
>Key is that this implies that we will provide the function, not that we
>will provide specifications of what the function should be.  Circles
>back onto the issue that OSDL CGL is not a distro, but that we provide
>guidance to those wishing to be distros.  The POC code is not intended
>to be the end-all, our specs are.
>That said, as an example, we have (picking 6.1.1, roll of the dice):
>"OSDL CGL shall provide the capability of configuring the scheduler to
>provide soft real time support so that the real time scheduling latency
>of a given task will not exceed a target offered by the vendor."
>Common point all over the document is the "OSDL CGL shall provide..." 
>Really, WE aren't providing anything.  We have a reference
>implementation, that will try to demonstrate that (most of) our
>requirements aren't unreasonable, and may even provide actual code for
>Linux distributors to use to meet some of the requirements when they
>ship releases that fulfill our requirements.  
>We do plan to provide a certification suite as well so that the
>distributors can show they've met our specifications.
>Some textual suggestions for the problematic clause:
> "An OSDL CGL-compliant distribution..."
> "The OSDL CGL reference implementation..."
> "OSDL CGL implementations shall provide..."
> "Implementations of OSDL CGL shall provide..."
> "OSDL CGL specifies that carrier grade Linux shall provide..."
>Geoff Gustafson triggered this, and provided the first two ideas.  It
>has been (strongly) suggested that we avoid the use of 'compliance'
>and/or 'certification' in the Requirements doc.  The second suggestion
>ties us too much to the reference implementation, but there is no
>requirement that a Requirement be met by using the same code that's in
>the reference implementation.  The rest are mine to try and work around
>these.  To say I'm underwhelmed by my creativity would be the
>overwhelming truth.
>Decision is mine to put forward, assuming we have something that
>achieves some sort of consensus.  Now that I'm sensitive to the current
>wording, I do not like it, but it may be the lesser of various evils if
>we don't have something succinct but useful.  I do like Geoff's first
>suggestion, but that does highlight the compliance/certification issue
>too clearly I think.
>Post or send me ideas, preferences, etc.  If I don't get any, I'll take
>my suggestion up with our roadmap coordinator for a command decision.
>Peter R. Badovinatz aka 'Wombat' -- IBM Linux Technology Center
>preferred: tabmowzo at us.ibm.com / alternate: wombat at us.ibm.com
>These are my opinions and absolutely not official opinions of IBM, Corp.
>cgl_discussion mailing list
>cgl_discussion at lists.osdl.org

More information about the cgl_discussion mailing list