AVR-Chat

--- In AVR-Chat@yahoogroups.com, Jim Wagner <wagnerj@...> wrote:
>
> I have a lot of respect for Jack Ganssle.
> This looks pretty complete.

I have a lot of respect for Jack Ganssle too.  He's been there and knows what it's like.  He's also not afraid to assign responsibility for most of this industry's immature practices where it belongs, with management.

However, this coding standard of his is just another bunch of opinions.  At risk of lecturing again, I would advise anyone reading it to be critical of his suggestions and take away from it only those ideas they can't find fault with.

Example 1 - On page 17 Jack specifies the "one true brace" style of forming blocks (compound statements).  This is 100% personal preference and there is absolutely nothing wrong with the other style of putting the opening brace on its own line (which I happen to prefer).

Example 2 - Also on page 17, Jack seems to be saying that nested 'if' statements should be replaced with a 'switch'.  Huh?  This is Jack being a writer (getting something out for publication) and not an engineer (thinking everything through).

Example 3 - On page 8 Jack tries to define 'module' and makes a mess of it.  He's close to defining a translation unit.  A more useful definition of a module is *any number* of source files that encapsulate some well-chosen group of data items and that exports functions to manipulate that data via a documented interface.  (Note that the data comes first, the functions second.  Jack has functions first and data in support of functions, which is backwards by modern thinking.)

Graham.