On Mon, 13 Dec 2010 11:14:19 -0800, Richard Clark
wrote:

On Mon, 13 Dec 2010 08:34:39 -0500, Registered User
wrote:

The phrase "80% antenna efficiency"
contains both a unitless number and dimensional metadata.

Actually, it doesn't - you are filling in the blanks with presumption.
Given that "efficiency" has been hijacked, the phrase above could as
easily relate to wind load.

Precisely why metadata is so important to describe data.The generic
term 'antenna efficiency' does not provide enough metadata to describe
the true meaning of the measure. Some additional metadata turns the
potentially ambiguous "80% antenna efficiency" into "the antenna's
radiated power divided by its input power results in a percentage of
80%"

80% survival rate at 100 mph. I could
well anticipate the counter-argument that "antenna efficiency" as you
intend it is power based, and I would counter-counter that power
delivered to the end user is far below 80% by 6 to 9 orders of
magnitude. There is nothing efficient about 79.999999% of the power
warming clouds and worms. HOWEVER, if the intended recipient receives
even that miniscule power with full quieting, then it is in fact 100%
efficient. In that regard, 80% and 100% as figures become meaningless
when they are both applied to the same statement of efficiency.

In this case
the metadata describes the dimensional calculation used to produce the
result. Remove the dimensional metadata and the phrase becomes "80%".

No, efficiency in engineering terms has always been well understood
and does not embrace this adornment.

There are many different expressions that calculate efficiency, so
wouldn't the adornment be the context aka metadata which describes the
expression used to calculate the result? If I write "80% efficiency"on
a piece of paper and hand it to an engineer, how can the engineer know
what is 80% efficient without any metadata being provided?

In an earlier posting I've seen
the distinction of miles-per-gallon and passenger-miles-per-gallon
stretched over the argument to fit it to efficiency. No, these two
comparisons (and what you largely characterize as the need to include
Metadata) are "figure of merit" measures. FOM is also dimensionless
but demands the Metadata you speak of.

The last sentence tells me you get it. 80% is a FOM with no UOM so for
the value to have meaning metadata must be provided.

FOM could easily lead you into a very energy (the engineering
consideration) inefficient solution (due to social or economic
considerations, eg. passenger-miles-per-gallon).


The whole purpose of data analysis is to aid in effective decision
making. Pretty much every decision requires some sort of trade-off.
By analyzing and reporting upon the expression results that different
combinations of dimension values provide, more information can be
provided the decision makers. The decision makers are the ones who
specify what data they wish to view (both dimensional data and the
results of multi-dimensional expressions) and the ways they wish to
view that data. All that information is placed in a multi-dimensional
cube. The decision makers can slice and dice the cube data into how
they want to view the data and determine relative FOMs for multiple
scenarios. Their questions are which scenario model is most suitable
and does the most suitable scenario provide a valid solution for the
problem.

AFA the passenger-miles-per-gallon calculation, the decision makers
could decide that more information is needed in terms of greater
granularity. Changing the granularity to stop-by-stop and providing
associated passenger and fuel consumption statistics will deliver more
information. The decision makers could also ask for data that
provides social and economic information. The multidimensional PMPG
expression could be changed to include the social and economic data.
More data and metadata provides more information.

The form of desired social and economic data and how it is to be used
in expressions is something the decision makers must determine. All
the information about required cube dimensions, attributes,
hierarchies, expressions, etc, are determined by the decision makers.
Those who build and load the structure have little say in the matter,
the decision makers decide what data and metadata is useful for their
analysis.

An image of a simple multi-dimensional cube can be seen at
http://books.google.com/books?id=AFC...page&q&f=false
or
http://tinyurl.com/337m42z

Each dimensional intersection uses the same multi-dimensional
expression (MDX) to provide a result. The UOM at each intersection is
the same (dollars) but the results differ because different sets of
dimensional arguments are used in the MDX. The amount $3,156,834 is
raw data but the cube tells us it represents total sales for the
Mythic World product in 2005. That metadata must be used to describe
the result if that value is to be used away from the cube.

Change that cube's dimensions and the values they contain to be
antenna-related, add a few more antenna-related dimensions and
attributes, create some expressions that use dimension values to
provide appropriate measures and the structure could be used for
antenna analysis.