reviewing best practices example in publishing section

[paolap]

I just checked a random example we give as best practice and there's not much info in the file and/or description. Notably the variables don't have standard_names. I think we should review these examples and make sure they really are best practice.

[hot007]

Is it my wave data? That is my favourite "dos and don'ts" dataset - a lot of wave variables didn't have standard names in the version of CF that was active when we originally prepared it in 2012. Should I retrospectively add them now, knowing that would change the data? I genuinely don't know. There's a bunch of things wrong with that dataset, but there are reasons (e.g. the errata about changing variable names halfway through), honestly everything I publish has problems!!! Happy to swap it out for better stuff, i'd just prefer not to use CMIP but rather regular research data as CMIP contributors will hopefully not be the main audience of this book.

[paolap]

Nope that's fine and we might actually use somehow your presentation to expand on why it is a good example?

https://dapds00.nci.org.au/thredds/dodsC/ux62/access-s2/reanalysis/ice/aice/di_aice_1981.nc.html
and a short description considering the complexity of the data, there's no link to extra documentation etc.: https://geonetwork.nci.org.au/geonetwork/srv/eng/catalog.search#/metadata/f9491_7340_7243_0479

I don't mean to say it's horrible but it's not best practice, it might have been at the time but not now.
And it's the only one I checked as it's the first but it occurred to me that these were added without having been checked. I'm fine with having a realistic example where some requirements are skipped because they aren't applicable, and I agree that showing CMIP is not a good example as there's a lot of requirements there which are meaningless to more common cases. It's more like we should actually review what we listed and see if we would consider now an attempt to best practice.