Re: The Last Word on Comments (was Re: Hungarian notation)

[kcline@sun132.spd.dsccc.com (Kevin Cline)]

In article <4pkkr6$t31@michp1.redstone.army.mil>,
Kevin J. Weise <kweise@c3i-ccmail.sed.redstone.army.mil> wrote:
 >Darius Blasbans <darius@phidani.be> wrote (with deletions):
 >>Clark L. Coleman wrote:
 >>>I shudder to think that
 >>> someone is paying them to develop code if they really think you add
 >>> comments after the code is written.
 >>
 >
 >If you think this practice is bad, wait until you hear/read this:
 >I worked for a company several years ago where part of the development
 >environment was a PDL generator.  You got it!
 >   They developed the code first, then ran it through a program to
 >   produce the PDL to put in the design document!!  And they believed
 >   it was a perfectly reasonable way to do things!!
 >
 >I can understand desiring to reduce the documentation effort, but to go
 >from B-specs to code using an undocumented process that back-filled the
 >design?  I can't wait to see responses to this one.

These practices are typically a response to documentation requirements
that result when the 2167A standard is not properly tailored.  This
can happen because the contractor doesn't know how to demonstrate that
efficient development practices are 2167A compliant, or because the
contracting officer doesn't understaind 2167A tailoring and covers his
ass by requiring every item of every DID in the document.  Probably
most of the verbage in the design documentats was useless for any practical
purpose, but supplied for some 2167A requirement.

The developers thought that writing PDL before coding was a waste of
time.  They were right.  For many projects, experienced developers can
proceed directly from OO design to API definitions (e.g. C++ class
definitions or Ada module specifications) and then to implementation.
Only DoD contractors working on cost-plus contracts have time to
write PDL, or even worry about it at all.
-- 
Kevin Cline