SIunits and quantity library concerns

From: Kevin Lynch (krlynch_at_[hidden])
Date: 2001-10-04 11:29:31

wb_at_[hidden] wrote:

From: wb_at_[hidden]

> Part of the reason, it seems to me, may be that there are some serious
> abuses of nomenclature in many of the postings.  Another reason, I
> speculate, is that no one has yet put forth a set of concepts that
> would cover the desiderata over and above what SIunits provides.

While I would agree with you to some extent on the former (we've
certainly been talking past each other as a group, but I suspect that
is because many of us are using different, yet completely valid,
nomenclatures), I would disagree with you on the latter.  But I'll
return to that later.

> The International System of Units (SI) very carefully distinguishes
>         a (measurable) quantity
> from
>        the units in which such a quantity can be measured.

I completely agree with you characterization of the nomenclature in the 
SI (for the interested, see the documents at 
http://physics.nist.gov/cuu/Units/bibliography.html 
for example); it is a very useful set of terms __in the context of the 
problem the SI attempts to solve__ I would disagree, however, that the 
terminology of the SI is useful for a more general discussion of units; 
it is too restrictive, with insufficient granularity to cover all of the 
cases that people here seem to be interested in for a
general framework of unit checking and conversions.  In other words,
the concepts of "quantity/unit" in general admits far more than do the
SI
concepts of "quantity/unit" (which are restricted to _physical_
quantities,
not admitting things like currencies and tags.  I think that _this_
issue 
of nomenclature is  one of the issues at the base of the lack of
consensus, 
not a misuse of  technical terminology laid out by the SI (although I
agree 
that such has occurred, and I am probably as guilty as others).

A further complication is that we all have a slightly different set of
ideas, from our different problem domains, as to what constitutes the
basis set of "fundamental measurable quantities" to which we attach
units: SI
provides one set (with, I believe, seven base units if I'm not
forgetting
any) and the "natural units" of particle physics require fewer (I'm not
going to 
get into the debate right now over whether there is only one
"fundamental
measurable quantity" or whether there are more....).  Derived units
provide a
whole mess of different issues (the cgs derived units for magnetic field 
strength and electric field strength are identical, which is not the
case in 
the SI).

As you noted later in your email, there are also issues related to
"tagged units", implicit conversions, commensuration, and such that do
not easily succumb to the SI framework, but nonetheless have a well
understood (if not well defined) meaning to the users of those
concepts.  For example, you ask

> On a related subject, it is also not at all clear to me what the
> correct concepts are in regard to such problems as "1 apple plus 2
> pears gives 3 pieces of fruit."  Suppose I subtract "3 pieces of fruit,
> take away 1 apple"?  Do I get "2 pieces of fruit" or "2 apples"?
> Both?  Neither?  

I would argue that the solutions to these issues are common sense, if
you
are willing to accept that unit conversions do not provide a bijective
mapping (that is, some conversions are not necessarily uniquely
reversible):
an apple "is-a" piece of fruit, but a piece of fruit is not an apple - 
units of apples can decay into units of fruit, but units of fruit can
not
be promoted to units of apples. So it makes sense to ask the following 
questions:
-> "How many pieces of fruit do I have if I have 2 apples and three
pears?"
-> "How many pieces of fruit do I have if I take 2 apples from my pile
of six pieces of fruit?"
-> "How many apples do I have if I have 2 apples and 3 pears?"
Clearly the answers are 5pof, 4pof, and 2apples.  The following
questions, however, do not make sense
-> "How many apples do I have if I take 2 pears from 6 pieces of
fruit?"
-> "How many square-apples do I have if I multiply 2 apples and 3
bananas?"
As in many things, I think the answers become more clear when you are
more precise with phrasing the question (this is the single biggest 
problem I have in my own research  :-).  And I believe that a general 
quantity/unit framework which supports and enforces the obvious
constraints 
is part of the wishlist.

> Finally, I am disappointed that no one has stepped forward to help
> address the oft-repeated request/need for a general-purpose Currency
> library.

I think that this is because many of us see this as a trivial
extension of a (currently nonexistent) extensible, general purpose
unit framework, and hence there is no need for a separate "currency
library" at this time; we're hoping that it will appear "magically" once
the
"hard part" has been done.

Let me say that, despite my disagreements on trivia here, I find that
SIUnits is a useful addition to the state of the art in the problem
domains that it addresses, and that once you've managed to solve the
"rational powers of units" issue, I would wholeheartedly support the
inclusion of SIUnits into Boost.  However, I still think that we
should keep searching for a consensus as to what is needed for a fully
general purpose library, and then figure out how to implement that
list.  
I will try to distill my general overview of the "wish list" that people 
have put together, and try to keep it updated somewhere as the
discussion 
continues.  I also have yet to find my discussion of units and unit 
systems that I wrote for my students a few years back, but I'll try to 
locate that and make it available as a potential starting point for 
discussion of a common terminology to ease communication.  If I can't
find
it, I'll write up something else.

-- 
-------------------------------------------------------------------------------
Kevin Lynch				voice:	 [phone]
Physics Department			Fax: [phone]
Boston University			office:	 PRB-565
590 Commonwealth Ave.			e-mail:	 krlynch_at_[hidden]
Boston, MA 02215 USA			http://physics.bu.edu/~krlynch
-------------------------------------------------------------------------------