Ken Mankoff

I've run into this issue too, and the xarray decision to use `float32` is causing problems. I recognize this is a generic floating-point representation issue, but it could be avoided...

This issue, based on its title and initial post, is fixed by PR #6851. The code to select dtype was already correct, but the outer function that called it had...

## Current algorithm ```python def _choose_float_dtype(dtype, has_offset): """Return a float dtype that can losslessly represent `dtype` values.""" # Keep float32 as-is. Upcast half-precision to single-precision, # because float16 is "intended...

- From: > This standard is more restrictive than the NUG with respect to the use of the scalefactor and addoffset attributes; ambiguities and precision problems related to data type...

## Packing Qs - If "the variable containing the packed data must be of type byte, short or int", how do we choose what size int? - What to do...

But... could I host the MeshLabJS code on http://example.com/ and the data on http://example.com/ ? In my example above, I assumed the bunny would be hosted on http://meshlabjs.net/

No. I think it has this ability, but may not be answered now that this ticket was marked as "closed". I'll re-open a new ticket where I phrase the question...

Hi. Perhaps a misunderstanding. I'm not trying to do cross-site loading, which seems to be a security issue. I'd just like to have a default mesh, hosted on the same...

A minor improvement would be to at least accept hostname instead of ip address for connection.

Even simpler than auto-connect, is allow hostname instead of IP address for connection.