I exported a correctly georeferenced drawing to dxf and it never became visible in the model domain of the groundwater modelling program where I imported it.

Reimport into Manifold showed a correctly georeferenced drawing.

Only after a support request to the groundwatermodelling guys did I look into the dxf-file with an editor and realized the data were exported with internal coordinates.

Why is my latest manifold edition exporting dxf - and presumably also shp - with internal coordinates?

Probably when creating your drawing it has adopted the local offset and scale parameters from an underlying image. Reproject your drawing in Manifold changing local offset to zero and local scale to 1. This has been a 'feature' of Manifold for as long as I can remember.

Colin,

indeed local offset and scale were inherited from the image. I created an upstream area from a surface (image).

Wouldn't you consider this a bug?

I followed your advice and in the assign projection dialog changed local offset and scale but all I got is local coordinates displayed in the status bar. Dxf export is still with local coordinates. And change projection - as suggested in an older thread - is not allowed.

No it's not a bug, just the way things are.

Right click the drawing component in the project pane and Assign Projection making no changes then rt click again and Change Projection making the changes I have noted. I don't see any reason that change projection of the individual drawing would not be allowed.

Not sure how you can recover from the steps you have taken so far and you might need to start again.

"Right click the drawing component in the project pane and Assign Projection making no changes then rt click again and Change Projection" - That's the instruction I needed - Works fine this way. Thanks!

indeed local offset and scale were inherited from the image. I created an upstream area from a surface (image).

Wouldn't you consider this a bug?

No, because usually when people are creating a new component based upon an existing component they want the coordinate systems to match exactly, in all particulars. There's considerable logic to that as well: whenever you create a new component you have to create it with some coordinate system in play. What should that be? One can imagine many different conventions to guide the creation of such a new component system, but one of the simplest and most obvious rules is that if it is created from some other component it should just inherit the coordinate system used by its parent.

If you want to use a different projection that's easy to arrange, but using the same projection as the default case is a significant convenience. It also has the side effect of reducing the likelihood of logic or performance errors for folks who are just getting started with coordinate systems. This helpful convenience always works without fail and is automatic and transparent for users (one less thing to think about) so long as you are working in a reasonably modern GIS system using reasonably modern formats that always keep track of the coordinate system information that's built into the data and used automatically by the GIS system.

All that goes out the window if one removes the built-in coordinate system intelligence by transferring the data into a non-geographic format like DXF. Now, because the format does not automatically keep track of coordinate system information, you have to take on that task. It's one of the hassles of using a non-geographic format to interoperate with geographic data, an example of why Manifold advises users to be wary of conveying GIS data through formats which fail to capture coordinate system information.

Keep in mind, by the way, that there is lots of data out there in DXF which must, absolutely must, retain the specified offset and scale for the data to make sense. The system therefore upon export sends the data out to DXF in exactly the form in which it exists without trying to second-guess the user's intentions.

Let's not leave out the fact that Manifold - seemingly uniquely - have introduced the "geotransform" for rasters into the projection metadata: this is "normally" kept separate. This is cunning, but very divorced from most expert raster users' experience. (IMO) What a lack of shouting stems from this, we seem to be Manifoldians by this measure alone.

You can edit your post for the next n minutes.

cuda shuda wuda

Well, it seems to me that by assigning image offsets and scales to drawings by default Manifold is second guessing users intentions and generally getting it wrong. It would save a lot of hassles if this was not default behaviour. Put something in the options dialogue to allow restoring it if there is really anyone who prefers this behaviour. We have continual problems with this despite being aware of the difficulties and emphasising it to everyone when teaching manifold.

I'm with you Mike, I continually get caught out by this despite being aware of it. I guess this is because I have never experienced this behavior with any other GIS and so never expect it.

I for one would welcome seeing it 'fixed' or changed.