Lee Corbin
>
> Colin writes
>
> > ACCURACY
> > Extent to which a measurement matches an international standard.
> >
> > REPEATABILITY
> > Extent to which a measurement matches its own prior measurement.
> >
> > For example the SICK DME 2000 laser distance measurement instrument
> > has an accuracy of about 10mm over 150m but a repeatability of 0.7mm
> >
> > Why does this matter?
> >
> > Because _within_ the measurement system is simply does not
> > matter what the accuracy is! As long as systematic errors are
> > repeatable, the systems behaviour will be repeatable...
>
> Sounds reasonable. And indeed, matches the *reliability* vs.
> *validity* of statistical measurements and performance. Does
> this distinction between accuracy and repeatability get the
> same kind of press that reliability vs. validity does?
>
I’m just talking about data sheets. Reliability is a 'mean time between failure' number in hours. 'Validity'? Dunno what that is. 'Availability'.... is another figure like reliability a %uptime, if you like, more to do with how much time is spent with the instrument out of service being calibrated.
When you buy instruments you have to understand the use to which it will be put. ‘Repeatability’ is an Ockham’s razor solution for instrumentation: it costs less! If the measurement’s accuracy matters outside the system being measured you must go for accuracy and get out your cheque book. All else equal it stands to reason that the cheaper option (repeatability) is what will be used by nature it will be selected.
> > So, for subjective experience: Yes it can be an illusion,
> > but a systematically erroneous, relentlessly repeatable
> > illusion driven by measurement of the natural world where
> > its errors are not important - .ie. not mission fatal to the
> > observer. Experiential qualities, in their solipsistic
> > presentation, need only be repeatable (my red/attached to
> > the linguistic token RED), not 'accurate' (internationally
> > standardized RED #12398765).
> >
> > This is equivalent to saying that the experience of HOT
> > and the actual hotness of reality (wobbly atoms) _do not
> > have to be intimately/directly related_!!! They can be
> > completely different and as long as the experience is
> > consistently used the behaviour of the experiencer will
> > be the same "OUCH".
>
> Well, wait a minute. The experience of HOT *does* have to
> be intimately related: otherwise, the machines we are would
> not have been built by evolution in this way. It serves an
> extremely important function for our survival as animals.
Here I’m afraid we have to disagree. What you are saying is that the experiential quality of HOT, which is _entirely_ generated in the brain (the meter output) from a sensory neuron or two (the measurement probe), is that our brain literally becomes hot? I don;t know about you, but to me very cold things feel like a burn. The same experiencing system is attached to different thermal behaviour in the world.
Also this is not what is found in any experimental tests. The transduction at your fingertip (in the candle flame) makes use of the thermal effects on cell excitability in your finger nerves. Thinking the way you suggest is like saying that the voltmeter indicates voltage by altering the voltage of the volt meter chassis, as opposed to altering the display. A fairly agricultural analogy but descriptive enough.
There is clear experimental evidence that the experiences do not match the physics of the real world. Take phantom limb! An amputee can have a feel a whole arm where there is none! That phantom experience of the arm is generated by brain material receiving pathological feeds from broken nerves.
We simply don't have to have the argment: it's over.
>
> > Haven't we all asked 'is my red the same as your red'?
> > Haven't we all concluded that we'd never be able to
> > ascertain the difference because it really does not
> > matter?
>
> No, only the philosophically inclined ever ask that. And
> yes, they conclude (or should conclude) that it doesn't
> matter and is actually a wrong question. It's analogously
> bad to "What is it like to be a bat?" another question
> that only a philosopher would ask, and which just derails
> thinking into unproductive channels IMO.
>
I was trying to instill an understanding of the implications of repeatability vs accuracy from the point of view of ‘being’ the instrument.
The last thing I need to do is get philosophical! :-)
> > ...we all point to the object and agree its red....
> > repeatability.... meanwhile the actual physical reality
> > of 'redness' is simply irrelevant and may not represent
> > any real quality of the observed system at all...
>
> That's *possible*, of course. Sometimes brains malfunction from
> the viewpoint of evolution. It was, after all, "actual physical
> reality of redness"
>
> ****WARNING WARNING WARNING PHILOSOPHICAL DANGER ALERT USE OF
> COLOR IN PHILOSOPHY EXCEEDINGLY DANGEROUS****
>
> okay, okay, It was, after all, properties of objects conveyed
> by the wavelengths of photons they reflected that gave a survival
> advantage to some species while other species (e.g. canines) found
> that information to be irrelevant to survival.
Yeah. Measure the right properties _for you_ and paint the world with paint made from it and you get to live. A brilliant programming technique. Brutal but it works!
>
> > I really wish mathematicians and philosophers and theoreticians
> > would get out and get dirty in the real world some times.....
> > half of the damned wordfest would disappear immediately.
>
> Hear! Hear!
>
> > Grumpy today.... sorry.
>
> You ain't seen 'nothin. Wait until you are in your late fifties, pal.
>
> Lee
<marge simpson chuckle> :-) col
Received on Wed Aug 17 2005 - 20:32:00 PDT
This archive was generated by hypermail 2.3.0
: Fri Feb 16 2018 - 13:20:11 PST