Stan,
My comments are not meant ONLY at your remarks, but perhaps they crystallized 
my thoughts.
 
What are we doing here?
 
I think we have jumped in details of phrasing when we haven't entirely agreed 
on what we want to accomplish.  If we seek to change the FPLA rapidly, I think 
we need to consider both the current FPLA law and the FTC agency rules that 
support it (both available on USMA site).  When those two are taken together, 
what is currently required/allowed/forbidden, and what do we seek to change?
 
I am not sure we want to change any of the detailed rules about what prefixes 
are allowed and when to use them.  If so, there needs to be agreement about 
what is "broken" and how to fix it, not just vote for these words or those 
words.
 
To go back to my idea of agreeing on intent, I will throw out a perfectly 
combustible straw man.  Feel free to burn him, but articulate your intent, 
don't bog down in details. Suggested intent, viewed against the totality of the 
current FPLA law and supporting FTC rules:
*For standard packages, we wish to change the requirement for net contents to 
include statement in Customary units to optional.  Where they are used, we do 
not wish to change any of the requirements on exactly how they are used.  The 
requirements for statement in SI units should not be changed.  This change 
could be effective IMMEDIATELY on passage.
*For random weight packages, we wish to change the requirement.  Currently 
Customary is required, and SI is permitted as supplemental.  We prefer to 
change to a requirement of SI, with Customary permitted as supplemental.  This 
will require changing scales in retail establishments and requires a delayed 
implementation date relative to the first point.
*When SI or Customary net contents are stated, they are to be stated in 
accordance with existing (FTC) rules.  It is not the intention to change any of 
the authorized prefixes, abbreviations, rule of 1000, etc.
 
Nobody has to agree to these ground rules.  Offer up debate.  But if we start 
on detailed wording, without groundrules, there is very little chance we will 
have agreement.  I think we need to back up a considerable distance and agree 
on ground rules first.
 
Stan: Your comments on significant figures are what we all learned in various 
science classes.  However, current FTC rules use a rule of 1000 and three 
digits max for expressing net contents.  However, they allow exceptions: two 
digits under 100 units, and suppression of "unnecessary" zeroes to the right of 
decimal.  We frankly don't know if a 2 L bottle is 2., 2.0, or 2.00 L.

--- On Fri, 4/10/09, Stan Jakuba <[email protected]> wrote:

From: Stan Jakuba <[email protected]>
Subject: [USMA:44509] Re: FPLA 2010
To: "U.S. Metric Association" <[email protected]>
Date: Friday, April 10, 2009, 9:07 AM

This is concerning two items:
1. The inclusion of the term significant digits in FPLA/FDA labeling; I believe
it has no place there.
2. The name of units.

The sig dig's: As we know, "the number of significant digits" is
a phrase that indicates several related issues. Foremost, it indicates how many
digits in a value can be verified by a measurement performed by the intended
instrument. In other words, the accuracy and precision of a measuring device
used to verify the last digit in a number will determine how many digits are
significant.

To illustrate, a value of 19.000 mm indicates 5 significant digits, which would
be appropriate if it reflects the measurement obtainable from a micrometer. A
value of 19.0 mm has three significant digits, which would be appropriate for
verification by a cheap calipers or high quality
tape measure. A value of 19 mm has two significant digits and it would be
appropriate for measurements in carpentry. And the number might appropriately be
rounded to one significant digit, 2 cm, if the information is intended to
reflect a rough measurement, a guess, or a situation where the ease of
remembering is more important than precision.

In calculated numbers where the accuracy of the verifying instrument is not
known, the number of significant digits in the result must not contain more
digits than was in the least precise number. (Precision reflects the no. of sig.
dig's and vice versa.) Similarly in conversions, the result must not
contain more sig. dig's than the original number but may contain fewer when
it is apparent from circumstances that an original number had exaggerated
precision.

Here is a refresher on significant digits and indeterminability (many had seen
this before).

There are three significant digits in a number like 77.8; five in 789.30;
one in 0.07; two in 0.071; four in 500.0, five in 10.302. The significance
is indeterminate in numbers such as 2000, or 20. Notice that prefixes make
it often easy to remove the indeterminability. Compare 2000 kW vs. 2 MW;
whereas the former is indeterminate (1, 2, 3, or 4 digits may be
significant, we do not know), the latter can be understood as more than 1.5
but less than 2.5. In calculations, the number with the fewest of sig.
dig's
determines the outcome. To illustrate, 17.5 m in 2 s is 9 m/s, whereas 17.5 m
in 2.00 s is 8.75 m/s.

Instruments used for the determination of the numbers on a nutritional label
can verify far more significant digits than the public needs to be shown.
Similarly, the ingredients - they could be specified more closely but it is
not needed. On the contrary, it is desirable that the numbers on labels are
as simple as possible which implies also the avoidance of the decimal marker
wherever practical. Thus many of  those numbers fall into the indeterminable
category, such as 300 mg,  350 mg, 500 cm³, 2070 kJ, etc.

Therefore, I would drop the word "significant" and suggest instead:

" ....select a prefix that will enable a value to be expresses with no
more that 3 or 4 digits. The 3 digits count is preferred. The 4-digits count is
to be used only if it is needed to retain the same prefix in a column of
numbers. For example, 1000, or 2070 with 870. In the 4-digits number there will
be no space or any other divider between the first and the second digit (counted
from the left)."

The above paragraphs suffices. There is no need to go into "significant
digits" in the FPLA/FDA regulation as it applies to labels on products on
supermarket shelves.

As a side issue, I see on many labels the unit in brackets with the quantity,
e.g.,
mass (mg) ..... 320.
I wonder if that ought to be considered although personally I am not in favor
of it despite the advantage it has in some circumstances.

Now the second issue. I believe the addition of names is a mistake. There is
not room for it on labels; every one label in my world-wide collection uses
symbols only; it is obvious that containers with nutrition labels are
exported/imported - not everyone speaks English, nor everyone uses the Latin
script/alphabet. Any imbecile knows those symbols by now. Why would we want to
introduce words where there was a simple and sufficient info in use from the
beginning of the labeling?

Here it is again as it was in Gene's original in the KIS principle.
(3) (A) (i) .......shall be expressed in g or its multiples;
          (ii)  ...........of linear measure, shall be expressed in m or its
multiples;
         (iii) ...... of area, shall be expressed in m² or its multiples;
        (iv)  .... of volume, shall be expressed in m³ or its multiples, or
ml, mL or l or L.

Stan Jakuba

Reply via email to