Missing Paris RGS prime meridian?

[mwtoews]

Paris (EPSG:8903) defined as 2.5969213 grad (×9/10=2.3372291699999996 deg) is very close to 2d20'14.025"E (according to GeodSolve the difference is 0.245 mm), so no real issue here.

Paris RGS (EPSG:8914) (RGS=Royal Geographical Society) defined as 2° 20' 13.95" E is 0.075" or approximately 1.529 m away from the Paris datum. According to the metadata, this datum was replaced by IGN in 1936 to the preferred EPSG:8903 Paris meridian.

Despite this note, should this meridian be added to PROJ as paris_rgs? Or ignored, as it's unlikely to be used?

[rouault]

I don't see a strong need to add that prime meridian as an hardcoded one. This must be used for old CRS only

[kbevers]

I don't see a strong need to add that prime meridian as an hardcoded one. This must be used for old CRS only

I had the same thought when looking at the equivalent issue for the Copenhagen meridian (#4030). Removing these hardcoded values would be a good thing to do when the time comes for a 10.0.0 release.

[mwtoews]

Removing hardcoded prime meridians from datums.cpp could break things. Should cs2cs -lm be deprecated? Or with an explanatory note that this list is incomplete and/or a legacy thing? Or should a prime meridian be dynamically loaded from proj.db, e.g.

cs2cs +proj=latlong +datum=WGS84 +to +proj=latlong +datum=WGS84 +pm=EPSG:8903

[rouault]

Removing hardcoded prime meridians from datums.cpp could break things.

We can't remove them. Or that would completely break PROJ.4 strings. I'd say let's live with the existing ones, and don't add more

[kbevers]

Or that would completely break PROJ.4 strings.

A guy can dream, right? ;)

Should cs2cs -lm be deprecated?

I think that would be a good starting point. Then remove the option in 10.0 as well as hide everything related to the old datums and prime meridians from the surface.