begin quoting Lan Barnes as of Sat, Dec 09, 2006 at 03:38:48PM -0800:
>
> On Sat, December 9, 2006 12:28 pm, Stewart Stremler wrote:
[chop]
> > I ran across an interesting article w/r/t XP and CMMI levels. The
^^^^^^^^^^^
> > claim is that XP is a vertical slice through levels 2 through 5:
>
> ... and the operant word is "claim." Note claiming source.
And you're claiming that it's level 0. So we have claims and counter-
claims, and no real discussion, just a bunch of rhetoric that amounts
to little more than a pissing match.
That being said, I'm not claiming he's right... just that it's
an interesting viewpoint. Pointing out that he's biased means
nothing, as nobody in this discussion has been a paragon of
objectivity and balance. Pot, meet kettle.
> I think the SEI would say that the CMM isn't a pu pu platter where you
> pick what you like.
Oddly enough, neither is XP.
Part of the problem is that like things aren't being compared. I'm
comparing XP-done-right against CMMI-done-wrong, and others are
comparing CMMI-done-right against XP-done-wrong. It's enjoyable as
a light-hearted religious war, I suppose, but it's not useful.
One of the more successful XP projects I know of was killed because
the XP team was making the rest of the senior engineering staff look
bad -- they were producing more working, debugged, and tested code in
a week than the senior engineers were producing in a month.
(And then I know of a project that claimed to do XP, but the programmers
were merely using an XP to engage in -- what did ya'll call it? -- seat
of the pants programming. The code was (is, actually) *terrible*, and
the design is almost completely unlike what the customer wanted. If
there were any of the XP-related artifacts around, I'd say that it would
be a great example of how XP fails... but I can't find any evidence that
the XP methodology was followed at all, aside from perhaps a little
pair programming.)
--
_ |\_
\|
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
