|
Posted by Ragi Burhum on August 26, 2004, 10:02 pm
Please log in for more thread options peruviangeek_deletethis@hotmail.com (Ragi Burhum) wrote in message
> > On Thu, 26 Aug 2004 00:39:01 +0200, "Erling Midtgard"
> >
> > >I am a master-student in archaeology at the University in Oslo, Norway. I
> > >have a little experience in using GIS, mainly ArcGIS and ArcView. Lately, I
> > >have been inspired by the concept of time-space geography as developed by
> > >Hägerstrand. The version I am thinking about using however is the one
> > >revised by Anthony Giddens.
> > >
> > >The conceptual basis of time-space geography, is that you have an x- and a
> > >y-axis representing the geographical space, like in a GIS. Then you add a
> > >z-axis which represent time. The thought behind is that actors (that is,
> > >persons), as well as societies, have a location (or more precisely a "path"
> > >or "trajectory") both in time and space - in the time-space. Time-space
> > >geography, then, could be applied to study both routinized, cyclical
> > >practices and historical change. Examples are day-to-day conduct of actors,
> > >or periodic (yearly or so) meetings of nomadic tribes.
> > >
> > >A crude model of a conceptual time-space might look like this (see attached
> > >file). This is a model of southern Europe (before mapping out the
> > >trajectories of any actors or societies).
> > >
> > >What I would like some input on is wheter it is possible using GIS to create
> > >anything like this, placing geographical layers along a z-axis, and
> > >representing it this way. I would also find it interesting to know wheter it
> > >is possible to use some extension (3D-analyst perhaps) to create a 3D model
> > >of the trajectories in the time-space, and how I would go about doing this.
> > >
> > >Regards,
> > >Erling Midtgard
> > >
> >
> >
> > The concept of temporal topology is embedded in in the ISO TC211 suite
> > of standards, and Dr Morishige Ota of Japan is actively promoting it.
> > It is possible that GML may become capable of encoding temporal
> > topology in due course - a paper showing how it could be implemented
> > has been submitted to the relevant standards bodies.
> >
> > The ISO TC211 web-site is at www.isotc211.org/
> >
> > I agree this is an exciting concept, and could even be essential in
> > the development of spatial databases - imagine the problems that arise
> > when allocating unique permanent ids to objects that then change with
> > time (e.g. a building). At present, a building is assigned an ID.
> > However, it is not clear what should happen when the building is
> > built, redeveloped or demolished . If objects are given a temporal
> > topology, then unique ids can be associated with the 4 dimensional
> > object, avoiding the difficulties that arise when working purely in
> > 3d.
> >
> > Paul Cooper
>
> - The Ocean's tide in a particular coastal region. The shoreline of a
> coast changes based on the time of the day. You could subdivide the
> shoreline between a high-tide, low-tide and "normal tide". However,
> everyday, the cycle repeats. For your particular problem, it might be
> the case that the shoreline has some of the non-spatial attributes
> being different (like maximum recorded wave height hitting the
> shoreline, but everything else being the same) in all three cases,
> except that the spatial properties (e.g. line location, length, etc)
> are different. A possible implementation would simply require two
> attached time fields; one marking the beginning, and another one
> marking the end of the particular tide cycle and the feature being
> replicated 3 times.
>
> -Population movement and changes. In this case, a cycle might not
> really be the best way of describing population movement/changes.
> Maybe a more proper way of representing your particular problem might
> be to include only a date/time field that records the *last known*
> point in Space/Time where certain knowledge of your population was
> believed to be correct. Every time that new knowledge about your
> features are recorded (like new population count), you would include
> those changes with a timestamp.
>
> Botch approaches are based on the idea that certain spatial attributes
> are valid on a certain point (or ranges of points) in time. Performing
> attribute queries, spatial queries or any other kind of queries would
> only involve adding a time parameter to your query.
>
> Visualization of point movement would also not be an issue, since many
> GIS software packages allow you "fill in the blanks" when you are
> trying to draw movement from point a to point b. Polygon visualization
> might be a different deal.
>
> To answer your question, yes you could to do this by using 3D analyst
> and/or Spatial Analyst. It might, however, require writing some custom
> code using ArcObjects yo suit your exact needs.
>
> Regards,
>
> - Ragi Y.Burhum
Great... my news client somehow truncated my first paragraph. Sorry
about the bad post.
Let me try a repost:
The research of temporal support withing a spatial database context
has vastly increased in the past few years. Personally, I have just
started reading some papers on the subject, so by no means I claim to
be an expert. To me, it is one of those concepts that the importance
seems so obvious, but yet nobody pays enough attention to it.
Currently, it is possible to do some basic modeling of Space-time
features in a GIS system. However, the way you want to express time
varies depending on the type of information you are trying to model.
Let me give you two concrete examples and simplistic explanation of
possible implementations
- The Ocean's tide in a particular coastal region. The shoreline of a
coast changes based on the time of the day. You could subdivide the
shoreline between a high-tide, low-tide and "normal tide". However,
everyday, the cycle repeats.
For your particular problem, it might be the case that the shoreline
has some of the non-spatial attributes being different (like maximum
recorded wave height hitting the shoreline, but everything else being
the same) in all three cases, except that the spatial properties (e.g.
line location, length, etc) are different. A possible implementation
would simply require two attached time fields; one marking the
beginning, and another one marking the end of the particular tide
cycle and the feature being replicaed 3 times.
-Population movement and changes. In this case, a cycle might not
really be the best way of describing population movement/changes.
Maybe a more proper way of representing your particular problem might
be to include only a date/time field that records the *last known*
point in Space/Time where certain knowledge of your population was
believed to be correct. Everytime that new knowlege about your
features are recorded (like new population count), you would include
those changes with a timestamp.
Botch approaches are based on the idea that certain spatial attributes
are valid on a certain point (or ranges of points) in time.
Performing attribute queries, spatial queries or any other kind of
queries would only involve adding a time parameter to your query.
Visualization of point movement would also not be an issue, since many
GIS software packages allow you "fill in the blanks" when you are
trying to draw movement from point a to point b. Polygon visualization
might be a different deal.
To answer your question, yes you could to do this by using 3D analyst
and/or Spatial Analyst. It might, however, require writing some custom
code using ArcObjects to suit your exact needs.
Regards,
- Ragi Y.Burhum
| 
XML Sitemap