[postgis-users] 2d VS 3d Point

[James Marca]

I agree.  I would go further.  Keep your geometries in a separate
table altogether, and link to the raw data with a join table.  Then
you can tag points with speed, timestamp, etc if you want, or not.
And you can combine point geometries into line geometries, and link
that table to the original data (one line to many data records) and
tag the line with timestamps and speeds.  Finally, in a time-space
prism, the speed is important, and so you'll find that you might want
to know altitude.  Walking uphill is slower than walking downhill,
which will warp your prism perpendicular to the altitude gradient.  

hope that helps.
James  
At approximately Thu, Apr 21, 2005 at 08:36:44AM -0700, Craig Miller wrote:
> I would keep time a separate field.  It isn't a spatial dimension.  Temporal
> analysis (time series, etc) and even spatio-temporal analysis typically have
> time as a separate parameter.  You may be looking for interactions between
> space/time, but they don't measure the same thing.
> 
> --Craig
> 
> 
> -----Original Message-----
> From: postgis-users-bounces at postgis.refractions.net
> [mailto:postgis-users-bounces at postgis.refractions.net] On Behalf Of Eli
> Dylan Lorimer
> Sent: Thursday, April 21, 2005 8:32 AM
> To: postgis-users at postgis.refractions.net
> Subject: [postgis-users] 2d VS 3d Point
> 
> Hola,
> I'm creating a table to store waypoints from a GPS. Ultimately, I'll be 
> creating a 3d visualization system based off of Hagerstrands space-time 
> cube - so time will be the 3rd dimension. My question is, in your 
> (anybody's) expert opinion, should I use a 2d point object to store the 
> lat/lon location and then keep time as a separate attribute 
> (aggregating them when needed) or store lat/lon/time as a 3d point 
> object which I can then map to a java class? Any thoughts?
> Cheers,
> ./dylan