Hi Frank,
Frank Schönheit - Sun Microsystems Germany wrote:
Hi Joerg,
We haven't identified the tests. The requirements was that they should
be "rock solid" and we have given those RESOURCE and MAIN FUNCTIONALITY
tests to customers and they were able to deal with them.
It makes sense to use firstly less than these ~45 and have a look if we
"stop the bleeding" [means many regressions on every MWS] or not. If not
we should enhance the test coverage. As I have written it is a
step-by-step process.
Sorry, that's the wrong order.
If we do not know whether those tests stop the bleeding, then we must
not introduce them as mandatory process elements. It should be the other
way round: We should have tests of which we know (or at least can be
pretty sure) that they will stop (most/lot of) the bleeding. *Then* we
can make them a mandatory part of the process.
I think we will never know, that a test-set will stop most of the
bleeding, when we don't try it. I expect that this will be an evolving
process. Start with a good guess, have a pilot phase. If the scenario
works reliable, make the test mandatory and then become better while the
process is running. When new bleeding occurs, consider to add according
mandatory tests.
Ciao,
Ingrid
Ciao
Frank
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
