mklotz said:
The fact that a bunch of engineers can't do arithmetic reliably in their chosen measurement system is hardly an indictment of the system. If it were, the Imperial system would be long gone. You're not much of an engineer if you don't do order-of-magnitude and sensibility checks on what you compute.
Marv,
That is the entire point of my rant. If I am working with stress in psi I
know that the answer I am looking for is going to end up being 1000's of psi (i.e. ksi). If I am working with stress in Pascals, my answer
could be valid as MPa or GPa or even (with some of the newer materials) TPa. There is no immediate feedback as to the order of magnitude value that sends me looking for errors or mistakes. This is as true for engineers who have never used anything
but the metric system (such as the French, German, and Japanese engineers I worked with while developing automotive airbag restraint systems) as it is for anyone else.
Arguments about the engineering uses of the system, e.g., thread standards, are equally hollow. Practical thread standards can be calculated and expressed in any measurement system. You're damning the system because of the use some humans have made of it. Measurement systems are not collections of engineering standards.
You bet your sweet @$$ I condemn uses where the
application fails to provide consistent results.
I built radomes for NATO many years ago (i.e. back when there were
five separate "standards" for metric threads). The system, built, qualified, and working fine in the Canadian north had to be
completely redone when we went to install it in Norway (which uses the DIN standard set) and
all the calculations redone to account for the different thread pitches and tolerances -- and redone
again when it came to French installation!
While things are somewhat better today with respect to major diameters and pitches, the
tolerances and allowances (which you need to know when applying critical loads to threaded connections) are
still based on varying standards. (This is one reason that the aerospace industry
everywhere uses Unified National fasteners.) Under current ISO practices and standards, a German made bolt has a 42% chance that it will
not mate with a Japanese made nut. You bet your sweet @$$ that this gets me up on my soap box screaming!
I do agree that the current metric practice of using only 10^(3*n) multiples is ill-conceived and we should return to the 10^n progression. Nevertheless, the advantage of the metric system concept lies in its internal consistency, the simple relations among units, the logical connection between weight and volume, and the explicit separation of force and mass.
And this is the
ultimate failure of the metric system! The unit of force, Newton, is so poorly meshed (1 N = .225 lbf) that the
kgf (kg-force) has begun to supersede it. I purchased a job-lot of 200 metric torque wrenches for NASA back in 1982. They came from France and had their calibration in
kg-m rather than
N-m. Nobody (but me) thought that this was strange (or wrong)! I just did a critical structure design for a Japanese company.
They insisted that all torques be specified in
kgf-cm.
When you go to the store in (say) France and purchase a kg of (say) cabbage, the
measurement they use when portioning out your cabbage measures
force -- not
mass! As a result, the
only advantage of the metric system (differentiation between force and mass) is
lost!
The "logical connection between weight and volume" is
density. The "problem" is that kg/m³ becomes a massively unwieldy number in a hurry. Analysis systems that are supposed to be metric (
SolidWorks,
Catia,
ProEngineer, etc.) have started using kg/cm³ over the past two decades. The metric
tables of densities use kg/m³ -- and the ease with which that value gets screwed up is amazing -- even for engineers who have
never used anything but the metric system (I see this all the time).
As I said several posts ago, metric lengths do not bother me in the slightest. Neither does Celsius temperature (though
I am still personally PO'ed at having bought into metric tooling at the time when the change was made to the base metre and Centigrade was replaced with Celsius). That's all well and good, but the
Newton and all the units that derive from it
suck as units of measure in the
real world -- and the proof of that is the prevalence of
kgf-derived units appearing to replace it -- even in "totally metric" countries!
One of the companies I have done a lot of work for coined the phrase, "There's the few and then there's Lew." I admit to the truth of that statement.
I design
mission critical components all the time. The strength of threaded connections is often the deciding factor as to whether or not a design works or fails. There are
nine theoretical combinations of size, tolerance, and allowance based factors that determine a Unified National thread series connection's properties. (In actual practice, there are five.) There are (4^5 =)
1024 actual combinations of size, tolerance, and allowance that must be considered when applying metric threads in a similar situation based on the "standards" as they sit today.
This is why
I say that the metric system is a set of
non-standards!
