Probably a silly question, but . . .

[VenusScott]

Is there a "best practice" for what the spatial reference should be for Open Spatial Data? Do I use what seems to be the "web" standard of WGS 1984 Web Mercator Auxiliary Sphere (3857) or our local/regional NAD 1983 HARN StatePlane Arizona Central FIPS 0202 Feet Intl? I've gone to some of the other Open Data site on the web and I've seen both. Is there an official opinion/standard?

[DanPatterson_Retired]

Document whatever you want to disseminate and don't assume that the end use will be the web in any event.  Don't forget, that 'standards' will differ over time and by location as your coordinate system aptly demonstrates...speaking as an NAD_1983_CSRS_MTM_9_2951... others with a more worldly perspective, like GCS_WGS_1984_4326 and WGS_1984_ World_Mercator_3395, may care little about the locale or local perspective.

[StanMcShinsky]

We have our data in a local coordinate system like NAD83 State Plane and only convert somethings for basemaps to Web Mercator in a separate FGDB.

-Stan

[PaulCrickard]

Most of ours is in Mercator when it goes out, but if you publish a service in ArcServer it doesn't matter because you can request any spatial reference you want returned in the outSR parameter and send any coordinates in using inSR.

[AndrewQuee]

Our open data is all in our local spatial reference for domestic use in desktop GIS/mapping.   However, we're getting a lot of use out of web services in WGS_1984_Web_Mercator_Auxiliary_Sphere_3857 because it goes with Google maps to make mashups, with clients consuming our vector data over Google's imagery.