C
On Tue, 09 Oct 2007 10:11:15 -0400, Peter Cole
<[email protected]> wrote:
[snip]
>No sarcasm, no condescension, just a frank opinion of his methods. We've
>discussed this at length before. Carl is always attempting to (dis)prove
>theory with actual measurements. The problem with this is it generally
>takes an even better grasp of theory just to set up the measurement. I
>don't mean to slam Carl, I appreciate the difficulty of what he's trying
>to do (I've worked in the precision measurement field) -- unfortunately,
>he often seems not to.
Dear Peter,
We're often at odds, but I think that you're doing a better job here
than I usually give you credit for.
In the same fashion, I don't mean to slam you. I appreciate that you
have considerable engineering skills and experience.
The problem that I see is that--and you're not alone in this--we on
RBT unfortunately tend to apply theory to a practical situation
without testing or even considering the possibility that our
theoretical, untested model may not include crucial real-world
factors, which by their very nature never how up in our purely
theoretical reasoning.
(We don't know what's wrong with our theories until reality rubs our
noses in our misunderstandings.)
Even more unfortunately, we then defend our theoretical position with
more theory and whatever ad hoc observations support us, which is more
debating than science.
(That's why we have the phrase "Back to the drawing board" to describe
what happens when theories are not tested.)
Here's about as plain and simple a bike example as I can think of,
where perfectly valid theory drew an obviously impractical conclusion:
http://groups.google.com/group/rec.bicycles.tech/msg/a14af7c4e0dc2fb0
When a closed chain is worn enough, you can easily invert it, despite
theory, turning it inside and out like a rubber band.
Similarly, I found to my surprise that the spoke tension on a cheap
MTB wheel dropped when I let the air out of the tire, instead of
rising as theory and even testing on 700c wheels showed it should do.
And I found that the contact patch refuses to expand or contract
nearly as much as RBT theory predicted when tire pressure is raised or
lowered. It turns out that a pressing on the outside of an inflated
canvas toroid is more complicated than a rigid steel piston rising or
falling in a rigid steel chamber.
My favorite non-bicycling example of theory blinding us to reality is
the Mpemba effect, where hot water freezes more quickly than cold
water, a result that still has no good theoretical explanation, but
which can be demonstrated with precision measurements--which aren't
actually necessary, since the effect is pronounced enough to be
well-known among practical folks who freeze things for a living.
Here the Mpemba effect is shown with precision measurements, the
bottom graph being hideously clear:
http://www.picotech.com/experiments/mpemba_effect/results.html
There are still web pages that insist that the Mpemba effect must be
due to poor test procedures, even though the original article
carefully eliminated the very objections raised, but more and more the
multiple explanations are complicated and uncertain:
http://hyperphysics.phy-astr.gsu.edu/hbase/thermo/freezhot.html#c3
http://en.wikipedia.org/wiki/Mpemba_effect
The real lesson of the Mpemba effect was that it took a visiting
professor to take Mpemba seriously. Instead of telling Mpemba not to
waste time asking why hot water froze faster than cold water, since it
was obviously theoretically impossible, the professor said let's see
what happens and discovered that this bit of myth and lore was quite
true:
http://physicsworld.com/cws/article/print/24493
A sound theory by definition leads to a sound conclusion, but in the
real world things are often so complicated that describing the
situation takes more sound theories than we first imagine.
My vote is for testing and measuring, particularly when there are
disagreements. Squeezing spoke pairs, for example, has lots of
plausible theory and plenty of enthusiasts behind it, but we still
have no tests confirming that it prevents broken spokes, much less
demonstrating how.
Cheers,
Carl Fogel
<[email protected]> wrote:
[snip]
>No sarcasm, no condescension, just a frank opinion of his methods. We've
>discussed this at length before. Carl is always attempting to (dis)prove
>theory with actual measurements. The problem with this is it generally
>takes an even better grasp of theory just to set up the measurement. I
>don't mean to slam Carl, I appreciate the difficulty of what he's trying
>to do (I've worked in the precision measurement field) -- unfortunately,
>he often seems not to.
Dear Peter,
We're often at odds, but I think that you're doing a better job here
than I usually give you credit for.
In the same fashion, I don't mean to slam you. I appreciate that you
have considerable engineering skills and experience.
The problem that I see is that--and you're not alone in this--we on
RBT unfortunately tend to apply theory to a practical situation
without testing or even considering the possibility that our
theoretical, untested model may not include crucial real-world
factors, which by their very nature never how up in our purely
theoretical reasoning.
(We don't know what's wrong with our theories until reality rubs our
noses in our misunderstandings.)
Even more unfortunately, we then defend our theoretical position with
more theory and whatever ad hoc observations support us, which is more
debating than science.
(That's why we have the phrase "Back to the drawing board" to describe
what happens when theories are not tested.)
Here's about as plain and simple a bike example as I can think of,
where perfectly valid theory drew an obviously impractical conclusion:
http://groups.google.com/group/rec.bicycles.tech/msg/a14af7c4e0dc2fb0
When a closed chain is worn enough, you can easily invert it, despite
theory, turning it inside and out like a rubber band.
Similarly, I found to my surprise that the spoke tension on a cheap
MTB wheel dropped when I let the air out of the tire, instead of
rising as theory and even testing on 700c wheels showed it should do.
And I found that the contact patch refuses to expand or contract
nearly as much as RBT theory predicted when tire pressure is raised or
lowered. It turns out that a pressing on the outside of an inflated
canvas toroid is more complicated than a rigid steel piston rising or
falling in a rigid steel chamber.
My favorite non-bicycling example of theory blinding us to reality is
the Mpemba effect, where hot water freezes more quickly than cold
water, a result that still has no good theoretical explanation, but
which can be demonstrated with precision measurements--which aren't
actually necessary, since the effect is pronounced enough to be
well-known among practical folks who freeze things for a living.
Here the Mpemba effect is shown with precision measurements, the
bottom graph being hideously clear:
http://www.picotech.com/experiments/mpemba_effect/results.html
There are still web pages that insist that the Mpemba effect must be
due to poor test procedures, even though the original article
carefully eliminated the very objections raised, but more and more the
multiple explanations are complicated and uncertain:
http://hyperphysics.phy-astr.gsu.edu/hbase/thermo/freezhot.html#c3
http://en.wikipedia.org/wiki/Mpemba_effect
The real lesson of the Mpemba effect was that it took a visiting
professor to take Mpemba seriously. Instead of telling Mpemba not to
waste time asking why hot water froze faster than cold water, since it
was obviously theoretically impossible, the professor said let's see
what happens and discovered that this bit of myth and lore was quite
true:
http://physicsworld.com/cws/article/print/24493
A sound theory by definition leads to a sound conclusion, but in the
real world things are often so complicated that describing the
situation takes more sound theories than we first imagine.
My vote is for testing and measuring, particularly when there are
disagreements. Squeezing spoke pairs, for example, has lots of
plausible theory and plenty of enthusiasts behind it, but we still
have no tests confirming that it prevents broken spokes, much less
demonstrating how.
Cheers,
Carl Fogel