[Geowanking] implicit wgs84 in geoweb coordinates until 2015 ?

[Allan Doyle]

Nice new thread, Mike. Interestingly, the more accurate our  
measurements become, the more we put ourselves in the position of  
having to deal with aging data. In an era of 10cm accuracy, phenomena  
such as continental drift become first-order disruptions in our  
existing data. What will NGA, NASA, ESA, etc. do with their petabytes  
of data as the world itself moves (non-linearly, no less) away from  
where it was when the data were collected?

There are a few ways this can lead us to think...

1. Don't use geocentric datums. Maybe the older location-based datums  
were not such a bad idea after all. Pegging a point on each  
continent, or maybe even smaller regions might be more stable in the  
long run. Then instead of having to compute new locations for every  
point in a dataset, you only have to deal with relative differences.

2. Use address-style locations. Forget about lat/lon. Use the address  
of something. Maybe we need finer-grained addressing schemes.

3. Don't worry about it unless you are a geodesist or operate a major  
government body. By the time the data you and I deal with becomes  
hopelessly out of whack, probably much later than 2015, quantum  
computing will be able to recalculate all the points in one fell swoop.

4. Don't worry about it and in 2015, there will be GPS units that  
provide WGS84 readouts much like they produce NAD27 or other datums  
today.

In reality, a combination of the above will keep us going. The micro- 
robotics people will use sensors to not fly into walls rather than  
rely on absolute positioning, FedEx will still be able to find my  
house, and the military has the ability to obliterate sufficiently  
large patches without having to really worry about a few cm here or  
there.

Perhaps the best strategy for locative use right now is to all agree  
to use WGS84 to show how good it can be when everyone agrees on  
something, and then be able to move ahead with wgs2015 when it happens.

Regardless, it's good to have both eyes open, so thanks for posting  
your notes here!

     Allan


On Oct 12, 2005, at 21:32, Mike Liebhold wrote:

> Following the long continuing  discussions about the wisdom of  
> automatically assuming wgs84 as the standard datum for describing  
> coordinates, particularly elevation,   in a prospective geoRSS or  
> geoRDF,  I made an effort today to shed some light on the  
> dimensions of the issue.
>
> Here are my notes from a long cordial conversation this morning  
> with David Doyle, Chief Geodesist for the National Geodetic Survey.  
> http://www.ngs.noaa.gov/ David is leading developement of a new  
> National Spatial Reference System,  a new,more accurate datum, as a  
> basis for US national surveys and mapping due to be completed in 2007.
>
> I explained to David that an informal group of programmers have  
> been talking about wgs84 as the implicit datum for a 'popular'  
> geocoding tools for web hypermedia.
>
> points:
>
> " wgs84 will be used for a long time, since it is embedded in US  
> GPS and DOD mapping."
>
> "Gallileo will not us wgs84 internally, rather something called  
> Gallileo Terrestrial Reference Frame"
>
> "The Russians are trying to revive their own GPS, which also uses  
> it's own datum."
>
> " Neither US GPS, nor Gallileo GPS can guarantee more than 2-3  
> meter accuracy. The number of decimal points is a meaningless  
> artifact."
>
> ( my aside: excepting of course, differential gps, and gps using US  
> network of CORS beacons - Continuously Operating Reference  
> Stations" see: http://www.ngs.noaa.gov/CORS/ )
>
> " When Gallieo goes online, dual channel GPS could yield one meter  
> accuracy"
>
> " In the future, commercial, multichannel GPS devices might convert  
> different gps channel outputs into a wgs84 format.)"
>
> " Tri-channel gps systems, US/EU/Ru might yeild sub meter  
> resolution by 2010 or so.
>
> " Around 2010-2012, the US plans something called GPS3, which could  
> yeild 10centimeter accuracy. which is required by a wide range of  
> precision applications including robotics, manufacturing, and fine- 
> grain virtual reality"
>
> "At that time there will be constellations of 70-90 satellites,  a  
> new Global Nav Sat System , GNSS , http://igscb.jpl.nasa.gov/ and  
> there will certainly be a new global spatial reference system,  
> administered by someone called  the International Earth Rotation  
> Service _the 800lb. Gorilla_" http://www.iers.org/ "
>
> "The new global reference will certainly supplant wgs84 as the  
> global standard datum by 2015"
>
>
> David also shared a few factoids:
>
> "55,000 USGS 7 1/2 minute topographic maps are encoded in nad27.  
> less than 1% has been upgraded to nad83"
>
> [ my aside, elevation data is obviated by new STRM terrain data.  
> other layers still problematic]
>
> "very high percentage of legacy Eurpean data is based on ED 50 datum"
> "east europe, asian" and other countrys all have different  
> coordinate systems"
>
> I agree that simplicity is necessary for a popular geocode format,  
> and the burden of importing strange data ought to fall on the  
> developer. And so, maybe it's OK until around 2015 to set a default  
> datum, based on wgs84 for users to geocode their own information,  
> but any generally useful mapping application, will have to include  
> a fairly robust capability to import,convert, and render common  
> external data. ( e. USGS nad27, or Euro ED 50, etc.)
>
> I am still very concerned, that wgs84 is problematic for a fine  
> grained gps for aplications like small scale robotics, sensor  
> networks, augmented reality..
>
> And while 2015 sound like a long time away today, remember that the  
> web is only 10 years old.  I'm grateful that html and http work  
> just fine today, probably than 10 years ago.
>
> Cheers-
>
> Mike
>
> Mike Liebhold
>
> iftf.org  |   starhill.us   |   starhill del.icio.us blend
>
>
>
>
>
>
>
>
>
>
>
>
> _______________________________________________
> Geowanking mailing list
> Geowanking at lists.burri.to
> http://lists.burri.to/mailman/listinfo/geowanking
>

-- 
Allan Doyle
+1.781.433.2695
adoyle at eogeo.org