can you do it with wkt strings instead of proj4 strings? We need to make sure that we can corectly parse what both ogc and esri call the values.

also you need to add Eckert_VI to the names array.

No sorry I don't have any WKT strings for this one.

Which names array? Isn't that what the projections.js file does?

I'm not sure I understand, it is added to the array on line 21. Is there a requirement for the underscore?

these are the links to the 6 wkt files, you don't need to remove the proj strings, we can test those as well

• http://spatialreference.org/ref/esri/53010/ogcwkt/
• http://spatialreference.org/ref/esri/53010/esriwkt/
• http://spatialreference.org/ref/esri/54010/ogcwkt/
• http://spatialreference.org/ref/esri/54010/esriwkt/
• http://spatialreference.org/ref/esri/53008/ogcwkt/
• http://spatialreference.org/ref/esri/53008/esriwkt/

yes the underscore is treated distinctly from a space

ok I'll change it, UTM needs to be fixed then because I copied the style form there.

you don't need to change it, you can have both in there with and without the space, I'm insisting on the underscore because that's how wkt describes it.

added the WKT strings but tests are failing with the same error as the proj strings before I added +datum=none as a param.

adding the following line to the end of the cleanWKT function in the wkt file gets the tests to pass, but it does so by turning off datums for all sinusudal and eck6 projections, will they never have datums ?

if(['eck6',"Eckert_VI","sinu","Sinusoidal"].indexOf(wkt.PROJECTION)>-1){
    wkt.datumCode = 'none';
}

so it looks like the general issue is just that sphere and none should be the same but aren't. Bear in mind that that I am not confident which of the two miller cylindrical version is correct

yes I think it's much safer to just set datum=none when looking for the "not_specified_based_on_authalic_sphere" rather than disabling datum transformations for that whole class of projections. There are probably other datum identifiers that should equate to datum=none and this filter can be expanded as required.

it looks like datum=none and datum=sphere might actually mean the same thing

What is still needed to implement this?

a refactor probably