Code layout and debugging time

[Jonathon McKitrick]

On Tue, Apr 22, 2003 at 01:12:32PM -0400, Lowell Gilbert wrote:
: Jonathon McKitrick <jcm at FreeBSD-uk.eu.org> writes:
: 
: > I was just reading an interesting statement in Code Complete:
: 
: Ah.  You are trying to indoctrinate yourself into Microsoft's ideas of
: good practices.  Thanks for warning us.

Well, it was given to me by a software manager here at work.  But I have
been modeling my code largely after style(9).  Frankly, I think my code is
far more readable as a result.

But in the process of surveying the source tree, as well as the rationale in
this MSFT press book, I thought it raised some interesting questions.

-----8<--------------
: > Doesn't this seem to contradict the idea that clear, well-formatted code
: > with lots of blank lines is easier to read and understand?  How could
: > debugging be any different?
: 
: No contradiction at all.  It just shows that the definition of "lots
: of blank lines" is somewhere below 16%.  Assuming we can trust the
: study (but it sounds about right to me).

Sixteen percent would mean every 6 lines or so.  That seems far too dense in
my opinion.  Even when you look at hardware drivers in the kernel, there are
often only 1 or 2 lines together, separated from the rest by comments and
whitespace.  I just don't get how debug time would 'increase dramatically.'


NOTE: Please CC me, as I am not currently subscribed.  Thanks.

jm
-- 
My other computer is your windows box.