Profile pic checks out

Agree - it would be very interesting to find out if our verbal communication sucks because we just didn't evolve a better way of syncing brain states, or because there fundamentally is no good way.

Unfortunately, is failing at it of course won't disprove the latter...

(And all of its protocols suck)

I believe when talking about humans, telepathy is usually called "speech"

But users should be able to expect that if they convert to ome-zarr with ilastik, they will get the best performance ilastik can offer them when they process those ome-zarrs.

While developing, my impression was that it's pretty good, but until we benchmark that, we don't really know :)

One remaining step is for the export (and reading) to really be optimised for ilastik's internal use. ilastik can't be the best converter for ome-zarr, it would have to be too configurable.

Of course this means power users can't optimise chunking, or compression. But for most users, the barrier of entry into working with ome-zarr is now zero.

I think that's what ilastik's main user base needs. Now our no-code users can seamlessly go from tiff, to h5, to ome-zarr, whichever best suits their machine.

This export format offers no options for configuration to the user, so necessarily it has to do everything right, or at least good enough, out of the box.

But I want to toot my own horn a bit about the fourth one 😁 ilastik now offers "multi-scale OME-Zarr" export.

This is the final step towards full OME-Zarr (v0.4) support.

Wow, good things can really take their time. Love seeing how it came together!

After reading the devs' issue regarding Windows support / not having an active team member running Windows, I've switched to micromamba...