Date:         Sat, 8 May 1999 14:14:16 +0000
Reply-To:     paulgill@jps.net
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Paul Gill <paulgill@JPS.NET>
Subject:      Deadlines and software development
Content-Type: text/plain; charset=us-ascii

Jules Bosch wrote (in response to Bob Fitz commentary at bottom):
 
>Speaking of software development...
>
>there is a book entitled, "Code Complete..."  My copy is in a client's
>office or I would provide the full title.  Anyway, a few years back it
>was recommended here, I think by Karsten Self.  It is just full of all
>kinds of information, tables, and lists, all to do with the
>"construction" of software projects.  Every time I refer to the book I
>am simply amazed to think that anyone could produce such an effective,
>detailed, voluminous book about such a problematic issue.
>
>The book was well worth the price.
 
Jules, the book you are referring to is: "Code Complete : A Practical
Handbook of Software Construction."  It was written  by Steve C
McConnell and is available from amazon.com for $28.00.  This book is a
classic.  Whether you program in SAS or any other procedural language,
this book is indispensable.  It has lots of great tips on naming
variables, documentation techniques, dangerous coding practices, writing
structured code etc.  The author has a nice writing style and the
frequent use of icons enhances the book. It's a 5 star book.
 
And on a related note, I believe Steve McConnell was a consultant at
Microsoft.  It is interesting because studies at AT&T demonstrate that
37.5 hours a week is the optimum for programmer productivity.  Their
data suggests that getting too far beyond this dramatically increases
the bug rate.  On the other hand, you have people at Microsoft routinely
working 60-70 hour weeks.  Who is right?  Well, one must remember that
Microsoft hires young, talented, extremely motivated individuals who
devote their careers to programming.  Many work long hours because they
want to, not because some manager says they have to.  And their code is
as good as it gets.  I think forcing people to work long hours to make
up for bad design decisions is counterproductive and results in
panic-mode code (i.e. bad code).  So, I believe it is your frame of mind
which determines whether longer hours can be productive or
counterproductive.  Also, I believe companies would be more productive
if they allowed much more flexibility in their work schedules.  As a
consultant, I can set my own hours, but permanent employees are often
incarcerated in the standard 8-5 (or 9-6, or 7-4) cell block, which can
be stifling to productivity. Everyone has a different biorhythm and
lifestyle preference that would better be served by dramatic changes in
work schedules.
 
Paul Gill
Director of Bioclinical Data Services Inc.
Author of  "The Next Step: Integrating the Software Lifecycle with SAS
Programming"
paulgill@jps.net
 
 
Bob Fitz wrote:
>
> In a message dated 8/05/99, 2:18:18, tchur@bigpond.com writes:
> <<SAS-L@UGA.CC.UGA.EDU>>
> Tom,
>
> I agree most wholeheartedly that while the SAS tech support in the USA is
> very good and better than most, the tech support here in Oz is the best of
> any computer related tech support that I have ever dealt with.   No bias
> here, I'm from the States but here on assignment.  Inspiration and
> perspiration are the key ingredients as long as the 50 to 60 hour a week
> spurts do not become a way of life.  If it takes 60 hours every week to do
> the job, there are one of two problems.  The first is that there needs to be
> another person helping or two you're not up to the task at hand.  massive
> hours in front of the screen produces bad code and burnout.
>
> Just my .02
>
> G'Day,
> Bob Fitz