[Ticket#: 2006040510000641] Re: [OT] Articles to read

[Richard Gaskin]

David Burgun wrote:

> One rule I try to always stick to, is that if there are bugs reported  
> in a release, however minor, they are fixed before the next major  
> release is made.

How many commercial products do you publish, and how large are they (KLOC)?

Of course we'd all like to aim for zero defects in our work, but in 
practice if a program is complex enough the developer will have to 
settle for less than being the only vendor to ship a program of that 
size bug-free.

HyperCard shipped with more than 500 known bugs, most OSes ship with 
thousands.  In terms of complexity Rev is somewhere between the two, a 
"virtual machine" of sorts, so we can expect some known bugs to persist.

With my WebMerge product I generally meet the goal of zero known defects 
at final release, and my support costs are a fraction of industry 
averages as a result.  But WebMerge is essential a state machine, far 
less dynamic than Rev.

With other products I manage we generally aim for zero defects only with 
issues that cause data loss, and evaluate the rest on a case-by-case 
basis in terms of customer value and company ROI.

I once did contract work on a project used internally at a Fortune 500 
business where their company politics compelled them to spend 15-30% 
more addressing bugs than the bugs were costing them.  Not a smart 
business move.  When I presented the ROI data to upper management they 
immediately put a halt to it, but the middle managers I had to deal with 
daily resented me from that day forward, since it cut the budget for 
their departmental fiefdom.

For more war stories on software quality vs. ROI, and a few good tips:
<http://www.google.com/search?q=good+enough+software>

-- 
  Richard Gaskin
  Fourth World Media Corporation
  Developer of WebMerge: Publish any database on any Web site
  ___________________________________________________________
  Ambassador at FourthWorld.com       http://www.FourthWorld.com