|
|
| GeoCommunity Mailing List |
| |
| Mailing List Archives |
| Subject: | RE: GISList: 80% of data having spatial reference? |
| Date: |
12/07/2001 10:34:42 AM |
| From: |
Jeffrey Stockhausen |
|
|
I agree, as well. As I see it, all data is spatial. Or rather, all
data can be spatial. It really comes down to how you look at it - the
reference that is applied. There is some data which appears not to be
spatial, but we have just yet to discover its true spatial nature. With
other data, we may have to just change out perspective. Now,
admittedly, the amount of magenta in the average football jumper does
not immediately jump out and scream spatial - but I think it could.
Now, everything is someplace at sometime. So, that in it self gives the
spatial context. It also gives the temporal context, which is and
should be intimate with the former. If I were to plot the locations of
all the players on a football team in x and y on a pitch (we'll ignore
time for the moment) and then added the attribute of colour (or amount
of magenta in their jerseys), would that make the magenta content
spatial? Well, maybe not. But that may be due to how we generally look
at things. Looking at data using Euclidean geometries is limiting, with
respect to both our ability to properly analyze and to manage the large
volumes being and having been collected. Spatial reasoning must be
liberated to allow the treatment of each attribute as a dimension in and
of itself. Multidimensional data contstucts allow such treatment,
implicitly encapsulating the important topological concepts of
neighborhood or regionalization. The basis of this approach to
(spatial) data handling is the Riemannian Hypercube Structure, which is
a complex coverage function where data elements may be of different
sizes, and the coverage may be in a number of dimensions - this from
1856 when Georg Bernhard Riemann defined multidimensional space in
terms of “Hypercubes” for his mentor Gauss in his paper, "Foundation of
Geometry" It gave us the Riemannian tensor matrix. Riemann found that
the relationship between the diagonal of the hypercube and the related
dimensions was quite simply an extension of the Pythagorean Theorem.
This establishes a geometric relationship between the dimensional
vectors in multidimensional space - and these vectors can be anything,
including that football player's magenta jersey. This concept is the
source HHCodes, and when implemented in such a way as to exceed 2D
encoding, allows for these relationships to be exploited. This issue
and related ones are a focus of much work being carried out by ISO/TC
211 with respect to schema for coverage geometry and functions.
Now, with respect to the issue of data collected and archived with a
distinct lack of component metadata, well, that is another kettle of
fish. Truly spatial data, in the traditional sense, that has been
intentionally or unintentionally severed from its inherent spatial
context will always be a problem. Unfortunately, some people never
learn.
Cheers!
Jeff Stockhausen
-------- Original Message --------
Subject: RE: GISList: 80% of data having spatial reference?
Date: Fri, 30 Nov 2001 08:31:13 -0500
From: "Anthony Quartararo" <ajq3@spatialnetworks.com>
Reply-To: <ajq3@spatialnetworks.com>
To: <gislist@geocomm.com>
I'd have to agree. Everything is spatial. In addition, it's not
static, so
by the time the exhaustive research is done to claim 80%, it probably is
something less or more. The more information is generated, the
percentage
is likely to decrease, at an exponential rate. This is both good and
bad:
good for geospatial service providers, bad because we'll never catch up
to
100% and there is probably an enormous global cost to not having spatial
reference to data. This cost gets higher as more data is created
(daily) at
an ever increasing rate. Thus people refer to "spatial" and
"non-spatial"
data when in fact, there is no such thing as "non-spatial" data, we just
don't know what it's spatial context is, yet.
Too, in order to claim ALL spatial data as having a reference, that
presumes
that ALL data has been inventoried and we all know how foolish it would
be
to make such a claim, so the research likely is statistical and
extrapolated, and without seeing the research or the algorithm for the
extrapolation, I would guess it is not current with the growth rate of
data,
data storage, etc.
Anthony
-----Original Message-----
From: McGuyer William C Contr 88 ABW/EMO
[mailto:William.McGuyer@wpafb.af.mil\
Sent: Friday, November 30, 2001 8:03 AM
To: gislist@geocomm.com
Subject: RE: GISList: 80% of data having spatial reference?
In all honesty is there any valid method of quantitatively determining
the
approximate percentage of ALL data having a spatial reference? In
reality
everything has some sort of spatial reference.
Bill McGuyer
GIS Consultant
Intergraph Mapping/GIS Solutions
(937) 257-2201, ext. 239
WPAFB 88 ABW/EM
-----Original Message-----
From: Bill Thoen [mailto:bthoen@gisnet.com&#
|
|
Sponsored by:
For information
regarding
advertising rates
Click Here!
|
