TenTec
[Top] [All Lists]

[TenTec] Re: Jupiter firmware problems...NOT! (SO!... )

To: <tentec@contesting.com>
Subject: [TenTec] Re: Jupiter firmware problems...NOT! (SO!... )
From: n9dg@yahoo.com (Duane Grotophorst)
Date: Mon, 2 Sep 2002 19:14:14 -0700 (PDT)
--- Robert & Linda McGraw K4TAX <RMcGraw@Blomand.Net>
wrote:

>   I don't really think this is the nature of all
> software.  I think it is the nature of Windows
> software.  Here are a couple of examples the I know
> of which have well written, reliable software sets.
> 
>   The Engine Control Modules (ECM) used by all of
> the auto makers are all microprocessor based, and
> their software does not appear to have the magnitude
> of bugs that windows software does.  If they did,
> the highways would be littered with the pistons,
> rods, etc, that came out of the engines when the
> software failed.

Absolutely true, but you have to wonder about the
modern day "motor heads" who are tweaking these
modules for various "performance" goals. You can bet
they see some bugs from time to time. But then they
are experimenting so a certain number of problems like
this are part of the game. 

>   Oh, I see your point.  Look at it this way, the
> ECM has sensor input from say 15 to 25 sensors. 
> Always the same 15 to 25.  Add a human intervention
> to the chain and you'd see bugs.  The sensors are
> known variables with set limits.  The human
> input..........well.  A different story.
 
And the ECM's purpose is completely and clearly
defined from the begining, our radios - assuming we
want new features modes etc. aren't so neatly defined.

 
>   The second example is the software that is used on
> the shuttle missions.  That software has a
> documented error rate that is something less than 5
> bugs per 1 million lines of source code.  I think
> the big difference here is the fact that the
> software engineers meet with the astronauts to
> design the software.  Everybody there seems to know
> that a software bug could mean that some of the
> people in the meeting may not be returning from the
> mission.  That is a rather large incentive to get it
> right the first (and only!) time.

If you want more examples of extremely reliable code,
and some that you may have already directly trusted
your life to, ponder the Airbus A319, A320, A330,
A340, plus the Boeing 777 series aircraft. They are
all 100% fly by wire. Issues with "bugs" causing
problems in any of those birds is virtually non
existent. And they are also designed to take direct
lightning strikes without upsetting the
hardware/software.
 
>   It's been suggested that Tentec offer updates and
> enhancement for say $20 to $50 per release.  Wonder
> if NASA astronauts would want to ride on a $50
> update.  I realize the systems are vastly different
> and more complex in the case with NASA, but the
> principle is the same.

I don't think such pricing per release is what most
are suggesting, but instead more of a yearly
subscription based approach. Bug fixes remain free,
but feature enhancements would come at a modest price.
 
>   My point is that software does not have to be done
> poorly.  It can be done so it is reliable.  I have
> never understood why the Information Technology
> community has not taken Microsoft to task over the
> quality issue.  The IT people are the ones who
> suffer the loss of productivity as they are always
> tracking down some new bug.

Or you could look at it another way, .... job security
for IT people. I've concluded some time ago that if
everybody were completely computer literate I wouldn't
be able to buy any Ten Tec radios.

>   You are correct in that software does not have to
> be done poorly.  As to Microsoft, business wise,
> it's a feature vs time vs cost issue.  It's a
> triangle any way you strech it.   It can be done so
> that it is reliable but again I enlist the 6 month
> rule.  Be first and take your chances.  Wait 6 mos
> and you get a reasonably solid package.
>   Bob

This trade-off equation essentially remains the same
in any software product with the exception of the
"your life and health depend on it" fields. Presumably
being amateur radio experimenters we should expect
some lumps and bumps. But you?re right, if you don't
want the lumps the "wait 6 months and see" strategy is
a good one.

Duane
N9DG


__________________________________________________
Do You Yahoo!?
Yahoo! Finance - Get real-time stock quotes
http://finance.yahoo.com

<Prev in Thread] Current Thread [Next in Thread>