Re: Request to Ken (RE: Coding Standards)

• Previous message: Universe: "Re: RBAC and "Role Hierarchies" ?"
• In reply to: Robert C. Martin: "Re: Request to Ken (RE: Coding Standards)"
• Next in thread: Robert C. Martin: "Re: Request to Ken (RE: Coding Standards)"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Date: Sat, 28 Aug 2004 22:41:23 -0400

Robert C. Martin <unclebob@objectmentor.com> wrote:

> On Thu, 26 Aug 2004 14:53:19 +0200, "Bjorn Reese"
> <breese@see.signature> wrote:
>
> >On Wed, 25 Aug 2004 16:13:01 -0500, Roger L. Cauvin wrote:
> >
> >> is the wiser mission-critical client, one who hires subcontractors of
> >> dubious competence from many different countries and insists on coding
> >> standards that legislate good programming practices, or one who spends extra
> >> money and time to interview and hire programmers that understand these
> >> practices from the start?

Enuff of the best aren't usually available at will.

> >First, clients hire subcontractors, subcontractors hire programmers.
> >Clients do spent time interviewing the subcontractors, and the
> >subcontractors do spent time interviewing (or selecting and
> >educating) the programmers, but when you work on project that
> >have hundreds of programmers there is a good chance that some
> >detail can get lost -- this detail could cause catastrophe. Hence,
> >coding standard as a safety device.
>
> I disagree. A coding standard in that environment is a security
> blanket. It provides the illusion of security without providing
> security itself. If the programmers aren't bright enough to know
> about the evaluation rules of && and ||, they shouldn't be programming
> on a mission critical project.

So langs using && and || are the only suitable ones for mission
critical projects?

PLUS:
        ~ Not all C/C++ coders are at the same level and what level is OK
for some areas is too low for others.
        ~ Some tricky code should often be discouraged regardless of level.

> >The point that I have been
> >trying to make throughout this discussion is that coding standards
> >are just one of many ways to raise quality, one of many safety
> >devices.
 
> Only if they are well reasoned and agreed to by the team. A coding
> standard that is legislated upon a team probably *reduces* safety.

Why the assumption standards are always imposed from above? Most of
the projects I've worked, the coders created the largest part of
standards.
 
> BTW, I think any project that has 100s of programmers needs to be
> rethought. The research evidence that shows that large teams are a
> strong predictor of failure is common and hard to refute.

Large teams:
        ~ without proper methodology and methodology tailoring is
problematic.
        ~ lacking adequate documentation is problematic.
        ~ lacking overall design plans is problematic
> If the
> programmers can be partitioned into smaller ... teams

Isn't it silly to assume that projects with 100's of coders *wouldn't*
be partitioned

> self organizing, teams

How do we self organize the proper number in the proper teams at the
proper time via "self-organization" primary over leadership based upon
taking into account client schedule. money and timetable?

How do we self organize the proper number in the proper teams at the
proper time via "self-organization" primary over leadership holistic
investigation of the domain and use cases?

Elliott

-- 
Theory Leads, Practice Verifies

• Previous message: Universe: "Re: RBAC and "Role Hierarchies" ?"
• In reply to: Robert C. Martin: "Re: Request to Ken (RE: Coding Standards)"
• Next in thread: Robert C. Martin: "Re: Request to Ken (RE: Coding Standards)"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]