[Proj] Running the cities example in the pdf documents gives different results

[Martin Desruisseaux]

Hello Dean

>From the peanuts gallery (I'm not familiar about this document and about
Proj.4 changes), the differences that you are observing (except for
London) are of the order of magnitude that I would expect for datum
shifts. For example NAD27 to NAD83 datum shift can be close to 80
metres. I remember that some older Proj.4 versions did not applied any
datum shift unless explicitly instructed to. Newer Proj.4 versions apply
automatically a datum shift, but I do not know which one.

For example in EPSG database 8.9, there is 85 different datum shifts
from NAD27 to WGS84 and 57 different datum shifts from NAD83 to WGS84 -
not even considering newer NAD23 realizations like NAD83(86),
NAD83(2007), NAD83(2011), etc. The datum shift to use depends on the
geographic area of interest (e.g. the values are different for Canada
than for USA). I do not know if there is a way for having Proj.4 to tell
us which datum shift parameters it selected. If the selection algorithm
evolutes, I guess it could be one possible reason for the difference
that you observe.

    Martin


Le 09/05/16 à 15:41, Dean Schulze a écrit :
> I ran the cities example using the inputs on pg. 4 of the docs at ftp://ftp.remotesensing.org/proj/OF90-284.pdf, but I got different results than shown.  Here are the results I got:
>
> $ proj +proj=poly -r cities.lat.lon.txt 
> # coordinates for a few cities
> -4887445.45	7318110.56 Boston, United States
> -5542376.59	6982834.25 New York, United States
> 171219.46	5415571.82 Paris, France
> 485343.33	5730932.66’w London, England
>
>
> The document linked above shows this:
>
> $ proj +proj=poly -r cities
> # coordinates for a few cities
> -4887590.49 7317961.48 Boston, United States
> -5542524.55 6982689.05 New York, United States
> 171224.94 5415352.81 Paris, France
> -8101.66 5707500.23 London, England
>
>
> Am I doing something wrong, or did something change from when the document was written?
>
> Thanks.