[erdmann]

I would also like to sing the praises of nip2. I use it constantly in my work with enormous images at the Rijksmuseum. It's backed by libvips, which means that it offers a delightful combination of insane speed and unlimited image size. The finishing touches of the 717 gigapixel image of the Night Watch were done in nip2. Also, the spreadsheet-like nature of the interface is excellent for designing, documenting, and repeating complex workflows. Cells contain images, and the reactive trickle-down computation lets you build up multi-step transformations and computations cell by cell. Any cell's contents can be replaced by a new image at will, so it's very easy to repeat an analysis or transformation for new source images.

John Cupitt (jcupitt here), the main developer of nip2 and libvips, is super helpful, responsive, generous, and patient.

They can pry nip2 from my cold, dead hands.

[kawsper]

I've reported many strange bugs to jcupitt and I can speak to his really helpful and patient nature.

My favorite one was when JPEG2000s that had an even width was being shown in greyscale on any mac app: Preview, Safari, Pixelmator, etc: https://github.com/libvips/ruby-vips/issues/345

All images with an odd width was working fine.

However the real bug was to do with chroma subsampling (and a lot of other technical stuff that I don't really understand): https://github.com/libvips/libvips/issues/2965

[jcupitt]

Thank you!

(the money is in the envelope on your desk)