On Wed, November 29, 2006 1:09 pm, Stewart Stremler wrote: > begin quoting Lan Barnes as of Wed, Nov 29, 2006 at 11:36:46AM -0800: >> >> On Tue, November 28, 2006 7:48 pm, Stewart Stremler wrote: >> > begin quoting Lan Barnes as of Tue, Nov 28, 2006 at 06:59:08PM -0800: >> >> This has all the code words for an SEI CMM Level 0 environment. >> > >> > Given my experience working on or with CMMI projects, that might not >> > be a bad thing. >> >> It's not a bad thing. It's a REALLY REALLY bad thing. > > Heh. > > At project[-2], deployment consisted of the 3 best programmers spending > 3 days 'making it work'. Classic CMM Level 0, IIRC. I got it down to 1 > person spending half-an-hour. Without creating any documents or forms. > Because I didn't generate any "processes", it was still, technically, > Level 0 (IIRC). > > The problem with metrics is that instead of designing a solution to > solve a problem, people end up eventually designing a solution to meet > the metric. Sure, they *say* that they're trying to solve a problem, > but there's intent, and then there's what actually happens. > >> > Still.... Heroic efforts? >> > >> > I don't do frantic anymore. It doesn't pay off in the long run. >> >> Things are messy enough at level 3.5, which is about where we are. > > If you're still solving the problem instead of trying to match a metric, > you're doing good. More than good; you're godlike in your restraint and > wisdom. > >> Level 0 is where I do my personal projects, and look where they turn >> out. > > I hear you. Most of my projects are "in progress [suspended]"... :-/ >
I assure you, no one is more skeptical of formalism for its own sake than I. And as you imply, I have seen as much abuse and stupidity in formailsm as I have help. But I've learned that, properly done, developing processes and following them is a godsend. I don't ever want to go back to the old way. -- Lan Barnes Tcl/Tk Enthusiast SCM Analyst Linux Guy Biodiesel Brewer -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
