Received-SPF: pass
 receiver=logan.com; client-ip=24.25.9.102; envelope-from=echristley@nc.rr.com
Message-ID: <4393A295.9000001@nc.rr.com>
Date: Sun, 04 Dec 2005 21:14:45 -0500
From: Ernest Christley <echristley@nc.rr.com>
User-Agent: Mozilla Thunderbird 1.0.2-6 (X11/20050513)
MIME-Version: 1.0
To: Rotary motors in aircraft <flyrotary@lancaironline.net>
Subject: Re: [FlyRotary] Re: Intersting flight
References: <list-863134@logan.com>
In-Reply-To: <list-863134@logan.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

al p wick wrote:

>On Sun, 4 Dec 2005 14:32:10 -0500 "Ed Anderson"
><eanderson@carolina.rr.com> writes:
>  
>
>>I think the old adage KISS goes a long ways - as you know, it is
>>    
>>
>possible to 
>  
>
>>decrease reliability (not to mention increasing weight and cost) by  
>>increasing redundancy pass a certain point - more parts to break.
>>    
>>
>
>Wow, I TOTALLY disagree with above statement. But note that I define
>"redundant" as having independent failure odds (same as logical "or"
>circuit). We always have risk reduction with redundancy. HUGE risk
>reduction, because you multiply the odds. So 1 circuit has 1 in 100 odds
>of failure. 2nd independent circuit jumps the odds to 1 in 10000! That's
>why they run two power leads to the ECM on OEM cars. Also multiple ground
>leads. Much much safer. 
>
>Perhaps you allude to cases where you add a second switch, but power has
>to flow thru both switches for circuit to operate. In that case, yes, you
>actually increase your risk. 
>

Al, I think Ed alludes to the case that an individual can't actually 
determine what the hell he has once a certain level of complexity has 
been reached.  Remember, most of us don't do failure analysis for a 
living.  I can analize a simple system for failure modes.  The simpler 
the bettter.  Add in ten or twelve switches, and I'll have to start 
guessing at what I've got, and will most likely settle for a "That looks 
good", and move on.

I do that all the time when developing software.  When given a new body 
of code to debug or modify, my first step is to go through and remove 
all of the error handling and exception checking.  Oh, they will go back 
in (I never actually save the file I cut all that stuff from), but it is 
just impossible to analyze something that is overly complex.

Simple systems allow us (the average builder) to analyze all the 
possible failure modes and build in appropriate safeguards.  Adding 
redundancy can complicate the system beyond our ability to analyze, at 
which point it becomes a serious liability (hidden failure modes).

Can we agree on this statement?
Redundancy only increases reliability to the point that the designer can 
still analyze said reliability.

-- 
This is by far the hardest lesson about freedom. It goes against
instinct, and morality, to just sit back and watch people make
mistakes. We want to help them, which means control them and their
decisions, but in doing so we actually hurt them (and ourselves)."