Martin Yeo

Would have prevented a lot of the problems on #6518

Another thing I forgot to note was a previous conversation with @stephenworsley, who is also supportive of the test results being human-readable, so that's another vote: - **Against** hashing arrays...

> The multi-dimensional array parsing could perhaps be offloaded to the `json` parse? That would just leave the formatting to be handled by something bespoke in Iris (i.e. an alternative...

> The only major downside of this is that we'd then need to "officially" support it. > > Without getting all meta, are you proposing that we also test out...

@bjlittle do you have any opinions on this?

> @trexfeathers Do you think this is more suited to be a GH Discussion rather than an issue? Your call 👍. I wasn't necessarily calling for comments so much as...

>However I took at look at your MANIFEST.in, where there is `recursive-include src *.py *.nc *.txt`, so it's a bit of a mystery why not. Since this was posted, @pp-mo...

>but my experience is limited You probably have more HTML under your belt than most of us!

When we make these documentation page(s), they could include an encouragement to engage in GitHub Discussions if anything needs changing/clarifying - don't want our decisions to be set in stone...

I hadn't thought much about collapsing a dataless `Cube` but it would be pretty cool. The scalar coordinates and cell methods are useful information in their own right, and don't...