2 messages....................................Johnnie
 
 
----------------------------------------------
 
 
 
>Date: Sat, 4 Jan 1997 00:43:21 -0600 (CST)
>From: [log in to unmask] (Dennis McClendon)
>Subject: Re: Scale...
 
 
Darius Bartlett raises an interesting question
 
>does the concept of scale only start to
>apply when I create hard copy output maps?...
>the statement of scale conveys much more information than only
>the number of map units that correspond to x number of ground
>units... it can also tell us a lot about the reliability and
>generalisation parameters of the data.
 
In the GIS world, I have sometimes heard the phrase "nominal scale" used to
describe the compilation detail of the original source, which in your
example would seem to be 1:1.  I don't subscribe to GIS-L, but it would be
interesting to hear their discussions on the subject.
 
In my files is a fascinating theoretical article from GIS World about 1990,
suggesting that GIS data should not be stored as typical lat/long or other
coordinates, but instead should be a number corresponding to a tile on a
tesselated theoretical globe.  As the data points carried more precision,
they would also have a longer length, thus communicating their reliability
at different scales.
 
To illustrate, let's use square tiles, numbered clockwise from 1 to 4.  At
1:250,000, a city shape might fill tile 3, so the data point representation
is "3."  But tile 3 can be divided into four tiles, so the airport might
have data point representation of "32."  The airport terminal area,
digitized at 1:25000, might be represented as "323."  A site plan of the
apron at 1:2000 might put the fuel tanks at position "3234."  A CAD survey
at 1:100 might put the filler cap at position "32341."  You see how this
differs from a lat/long position, which might use the same coordinates
*carried to the same number of decimal places* for both city and filler
cap, if they came from surveys done at wildly different scales.
 
This process can be done on a spherical surface by using triangles or
hexagons.  I was intrigued by the ingenuity of this idea, and couldn't
resist passing it along to MAPS-L.
 
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
Dennis McClendon, Chicago CartoGraphics      [log in to unmask]
 
 
 
----------------------------------------------------------------
 
 
 
>Date: Fri, 03 Jan 97 14:54:27 +0800
>From: [log in to unmask]
>Subject: Darius' scale
 
 
Happy New Year, Darius --  I knew there'd be something interesting on the
digest today (the other ones I read had incredibly boring stuff today).
 
I think the concept of scale only becomes involved when you "represent"
the information.  For example, if I take the list of points from the data
logger the guy next to me is using, those are in meters or feet according
to how you set up the data logger.  So, I would say there is no scale
involved there (unless it's 1:1) because I'm not changing any of the
measurements to represent some other measurement.
 
If I decide to draw a map at something other than the actual size of
something, the scale could be 1:1000 or whatever, or even 1:0.5 if I draw
it twice the size recorded in the data logger.
 
How does that grab you?
 
vh
 
P.S.  Happy New Year!
 
\ /  Virginia R. Hetrick
 O   Bellnet:  310.206.7588
 0o  Internet:  [log in to unmask]