it seems the thread is returned to its origin which for me is location measurement metadata (optical attributes to an open view)
regarding the idea that attributing a location measurement to an area grid is: >a non-reversible transformation > to your initial point data is one way of seeing it but I prefer assigning what timestamped measurement was taken by what instrument with what configuration. what's measured encompasses more than the point of measurement; defining area of measurement is very helpful. so reversibility is not that important because the entire measurment event is recorded with metadata. > > So either you have a point location, or a circular area, and you are > applying a transformation to change that point, or circle, into a > triangle. And you can't reverse that transformation, plus you don't > have any real certainty that your actual position is within the > resulting triangle. shape is less importatnt that minimum boudaries necessary to define area - RF is a funny thing, ya know - our location fixes may or may not be GPS which may or may not be circular >You want to be able to analyze and present your data by > defined areas... yes and absolutely. >But nor does your method of storing data need to be how you present the data. correct but a hierarchical scheme originating from a recursive function can be very database friendly, open and easy to understand. great thoughts, Rich - thanks - brian > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] Behalf Of Rich Gibson > Sent: Friday, June 15, 2007 1:52 PM > To: [email protected] > Subject: Re: New coding schemes -was [Geowanking] Open Street View > > > On 6/15/07, brian grant <[EMAIL PROTECTED]> wrote: > > > What is 'contemptible' about lat/lon? > > > > it's lousy for assigning attributes to an area; more than adequate for > > determining what area is under measurement. > > It occurs to me that you are looking for an alternate system because > you are not using tools that understand geometries. Take a look at > the PostGIS documentation dicusion of GIS objects: > http://postgis.refractions.net/docs/ch04.html#RefObject > > > > > I am still confused as to what is wrong with tracking the location of > > > your sensor notes using lat/lon. What is the problem with lat/lon? > > > > identifying their location with lat/lon is OK - it's storing > the resulting > > data attributable to a defined area where lat/lon becomes troublesome. > > I don't think you want to store your data 'attributable to a defined > area.' You want to be able to analyze and present your data by > defined areas... > > > > > > What is so attractive to you about losing resolution? > > > > I'm not losing resolution. in fact, I'm incorporating the > precision of the > > location measurement into the area defined. > > huh? It looks like you are applying a non-reversible transformation > to your initial point data. Arguably, because of GPS accuracy, your > original point data is really an area of probability-you have a 95% > (or whatever) chance that your true position is within x meters of the > GPS reported position... > > So either you have a point location, or a circular area, and you are > applying a transformation to change that point, or circle, into a > triangle. And you can't reverse that transformation, plus you don't > have any real certainty that your actual position is within the > resulting triangle. > > > > > also, the mesh I'm suggesting has a direct correlation to lat/lon. > > it has to - it's how we measure location. > > > > my point is that our method for measurement doesn't have to be > how we store > > data. > > But nor does your method of storing data need to be how you > present the data. > > cheers, > Rich > _______________________________________________ > Geowanking mailing list > [email protected] > http://lists.burri.to/mailman/listinfo/geowanking > _______________________________________________ Geowanking mailing list [email protected] http://lists.burri.to/mailman/listinfo/geowanking
