Michel, Great Wikipedia link! I've found that most people get a glazed look in their eyes if I say precision is about repeatability, so I usually just stick with the number-of-digits explanation, or use the dartboard/target example if they show any interest at all.
Coming up with a way to state the accuracy so most people will understand can be a challange. One way is to say "within 20 meters at 2 sigma", another (less rigorous) way is "suitable for mapping at 1:20,000". And finally, the rather ambiguous "+/- 20 meters". And while these are similar, they are not really the same. But on the other hand, they're close enough for non-rigorous discussions... But anyway, I wanted to get people thinking about mixing 1:50k CanVec data (+/- 50m) with GPS-gathered data (+/- 5m). Is tagging with the "source" of the data enough? Thanks! Brent Fraser > 2008/12/10 Matt Wilkie <[EMAIL PROTECTED]> > >> Hi Brent, thanks for bringing up the oft-overlooked and even more often >> misunderstood confusion of accuracy over precision. >> >> > Most (some? all?) of the CanVec data originally came from the >> > 1:50,000 NTS topographic maps. >> > > Precision and accuracy is a subject that comes back all the time in > Geomatics. Thanks to > Wikipedia<http://en.wikipedia.org/wiki/Accuracy_and_precision>to help > me explain why the CanVec, Geobase, NTDB documents are rigth when > they describe the metadata. Precision as defined by Brent is realted to > the > computer world not "scientific" world. Precision is about repetition > measures not the number of decimals. > > Michel > > >> >> For clarity: which had their coordinates rounded off to the nearest >> metre (in UTM coordinate space), as per the NTDB specification. >> >> >> matt wilkie >> -------------------------------------------- >> Geographic Information, >> Information Management and Technology, >> Yukon Department of Environment >> 10 Burns Road * Whitehorse, Yukon * Y1A 4Y9 >> 867-667-8133 Tel * 867-393-7003 Fax >> http://environmentyukon.gov.yk.ca/geomatics/ >> -------------------------------------------- >> >> >> Brent Fraser wrote: >> > I'm sure they mean accuracy instead of precision. Precision is just >> > the number of digits stored/displayed, whereas accuracy is how well >> > the data reflects reality. Just because you chose to display >> > coordinates to the nanometer doesn't mean they are that accurate. >> > Not that I want to confuse the issue, but it can be important. >> > >> > Most (some? all?) of the CanVec data originally came from the >> > 1:50,000 NTS topographic maps . Within the past few decades some >> > have been updated from medium-resolution satellite imagery, and some >> > have been updated with data from the various Provincial 1:20,000 >> > mapping initiatives. At any rate, the "Quantitative Horizontal >> > Accuracy Value" is given in the metadata for each NTS sheet, with a >> > number 30 meters being common. >> > >> > To stir the pot even more, the Manitoba government >> > (https://mli2.gov.mb.ca//) has it's 1:20k topographic maps available >> > for free (and the license looks libre too). Their metadata gives >> > accuracy values (Horizontal_Positional_Accuracy) of 1.25 meters, 2.5 >> > meters, etc. Wow! I expected 20 meters at best. >> > >> > And don't get me started on the accuracy of hand-held >> > navigation-grade GPS receivers... >> > >> > Best Regards, Brent Fraser >> > >> > >> > >> > >> > >> > Richard Weait wrote: >> >> Hi folks, >> >> >> >>> From the good folks at Natural Resources Canada (GeoBase). >> >> >> >> "Yes you can use the data found on GeoGratis site . The licences >> >> are identical. The only differences are the copyrights, one is >> >> GeoBase, the other one is NRCan (GeoGratis). >> >> >> >> Please note: The data found on GeoGratis could have different >> >> planimetric precision and could not fit exactly with the precise >> >> GeoBASe data or OSM data. Please refer to the metada info of the >> >> files you will be using." >> >> >> >> I'm sure that we are all excited about the additional data. Please >> >> note the guidance regarding precision. >> >> >> >> Best regards, Richard >> >> >> >> >> >> _______________________________________________ Talk-ca mailing >> >> list [email protected] >> >> http://lists.openstreetmap.org/listinfo/talk-ca >> >> >> > >> > _______________________________________________ Talk-ca mailing list >> > [email protected] >> > http://lists.openstreetmap.org/listinfo/talk-ca >> > >> >> _______________________________________________ >> Talk-ca mailing list >> [email protected] >> http://lists.openstreetmap.org/listinfo/talk-ca >> > > _______________________________________________ Talk-ca mailing list [email protected] http://lists.openstreetmap.org/listinfo/talk-ca
