Russell King - ARM Linux wrote:
> 
> On Tue, Apr 03, 2001 at 07:04:02PM -0400, John Dorsey wrote:
> > One of our Computer Engineering professors is very much into software
> > robustness. He asked me to let him know if I came across any evidence
> > that this is not an April Fool.
> >
> > I'll believe that it's from an ARM Linux machine, and I'd buy that the
> > MTD drivers pulled the trigger. (But I'd be less likely to believe that
> > anyone would run the MTD code on a life-critical system.) There really
> > is an apfbioelectronics.com.
> 
> Well, its certainly a well-designed April fool, and going by the fact that
> the person hasn't replied to it yet...
> 
> However, the things are that:
> 
> 1. they are using a StrongARM, manufacturered by Intel.
> 2. Intel don't license their products for use in life-critical situations.
>    (read the small print at the start of any of their manuals).
> 3. A pacemaker is a life-critical situation.
> 4. We don't make any comments about Linux being the _most_ reliable OS out
>    there.
> 5. Linux is very heavy weight for a pacemaker, which has only one function.

        I disagree, I don't think it's so well designed. An april fools joke
should be either so believable to the target audience that it might just
slip by, or so contrived as to be immediately recognizable as humor by
anyone who's actually thinking. This is neither. Unless they're
appealing to the fear that somewhere someone really is that stupid. 

        What seemed odd to me was that I can't concieve of using an operating
system of any kind in a pacemaker, and wonder how anyone else can. Even
setting aside concerns of bulky hardware and relatively high power
drain. 

        The really fancy pacemakers these days actually increase the heart rate
with rising exertion, which is a lot more advanced than the classic
style, but when you think about it, it's almost exactly like the engine
controller in any late model car. 

        The way engine controllers work is, they have an eprom which contains a
3d map of what those who tinker with them call "performance curves" -
this is a very simple map of data, essentially a lookup table. When A
and B are equal to X and Y, set Z to N. 

        Currently the auto industry uses very low end microcontrollers to
handle this, and probably custom codes the software to do it. But i
don't see a reason why you couldn't implement the entire thing in an
ASIC. The auto industry probably sticks to microcontrollers so that they
can adapt one design to go into several different models, but there
aren't so many models of respiration in animals.  

        No need for an operating system at all, unless you'd prefer to hire
bargian basement software jockeys right out of college (or import them
from pakistan, or wherever they're cheap these days) rather than pay big
bucks for people who know what they're doing when they design a
pacemaker. Which is, these days, frighteningly believable. 

 - Eric

_______________________________________________
http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm
Please visit the above address for information on this list.

Reply via email to